Day 1: First Flight - Sphero Logic Basics

Robotics Engineer Pathway - Coordinate Systems and Movement

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Welcome. For the next 6 weeks you will work through 5 different labs - robots, drones, medicine, renewable energy, and a research showcase. Real experiments, real data. Today is Day 1.

The Hook: A Mars rover got stuck in a Martian sand dune. Engineers had 90 minutes to talk it out before sunset killed the radio. They could not see what was around it. They could only send numbers: heading, distance, speed. The rover did exactly what those numbers said. If the numbers were wrong, the rover was lost. Today you are those engineers. Your Sphero is the rover. Your code is the radio call.

NASA: 7 Minutes of Terror

Watch NASA engineers describe the math, code, and timing it takes to land a rover on Mars. You are the same kind of engineer today, just at a smaller scale.

Foundations - Heading, Speed, Duration

A Sphero BOLT takes three numbers to move: heading (0 to 359 degrees, where 0 is forward), speed (0 to 255), and duration (in seconds). Three numbers, every move.

Compass Cheat Sheet: 0 degrees = forward (away from you when aimed) 90 degrees = right 180 degrees = backward 270 degrees = left The blue tail light shows the back. Always aim with the tail toward you before you press Run.

1Talk to your partner: if a rover is given heading 90, speed 100, duration 2 seconds, where does it end up relative to the start?

2Predict: what would change if you doubled the duration? What would change if you doubled the speed?

I can name the three numbers a Sphero needs to move

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 Sphero BOLT, 1 iPad or Chromebook with Sphero Edu app installed, masking tape arena (1 m x 1 m), notebook, pencil. Charging dock central. Do not power on until told.

Today's task with Sphero Edu: program your robot to drive a perfect square - the same logic Mars rovers use to map a region of the planet. You will tune speed, heading, and time until the path closes back on itself. Code Here - Sphero Edu: Pick the device in front of you (works on iPhone, iPad, Android, Chromebook, Windows, macOS, or any Chrome browser): - Web app: https://edu.sphero.com/code - All-platforms download page: https://edu.sphero.com/downloads - App stores: search 'Sphero Edu' Sign in: tap 'Join Your Class' and enter the 6-character class code your instructor shares. Connect your robot: hold your BOLT near your device and tap its name when it pops up. The app finds it via Bluetooth.

This is Sphero BOLT

Carefully watch this video to learn the tool before you use it. Sphero's official 90-second intro to the BOLT robot. Watch this once before you connect.

Connect Sphero to the Sphero Edu App

Watch this video to see exactly how the technique works before you try it. How to pair your BOLT with the app over Bluetooth. Step-by-step.

The Hypothesis

1Write your predicted code in your notebook: four roll commands with heading, speed, and duration for each side. Use speed 60 to start.

The Build - Square Path

Iterative Design Rule: Your first run will not be perfect. That is normal. Real engineers do not get the answer on attempt 1. They get it on attempt 4 or 7. Today's goal is NOT a perfect square; it is to get within 5 cm of start by adjusting your code each round. Try, measure, adjust, try again. Repeat until it works. That loop IS engineering.

Watch the floor type. Carpet kills speed. Smooth tile lets the Sphero overshoot. The same code gives different answers on different surfaces - this matters for the rest of the week.

2Open Sphero Edu. Tap Programs. Tap +. Choose Block Canvas. Connect to your BOLT (the name shows on its underside).

3Aim the robot: tail toward you, blue light visible. This sets heading 0 = away from you.

4Drag four Roll blocks. Set heading 0, 90, 180, 270 in that order. Speed 60. Duration 2 seconds each. Add a Stop block at the end.

5Press Play. Watch where it lands. Did it return to start? Measure the gap with a ruler. Record in cm.

6Iterate: if it overshot, lower duration. If it undershot, raise duration. Run again. Record the new gap.

7Goal: get within 5 cm of start in 3 attempts. When you hit it, take a photo of your code with your iPad and paste the screenshot into your notebook (or sketch the blocks).

I drove a square path within 5 cm of the start

I recorded the surface type in my notebook

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

Get to Know BOLT+: STEM, CS & AI Learning

LONG VIDEO (skim, do not watch all the way through). Sphero's overview of how BOLT+ teaches AI. Watch 2-3 minutes to get familiar with the AI Assistant - that is the feature you are about to use to audit your triangle code.

Sphero Session: AI Literacy Through Hands-On Learning

LONG VIDEO (skim, do not watch all the way through). Sphero educators walk through the AI Assistant in real classrooms. Skim 2-3 minutes for ideas. Look for: 'Explain My Program' and 'Code Review' - both are buttons inside Sphero Edu.

Bring in the AI Assistant

The Sphero Edu app has an AI Assistant built in. It can generate code from plain English. It is fast. It is confident. It is also wrong about a third of the time. Your job is not to use it - your job is to AUDIT it.

1Tap the sparkle icon at the top of your canvas to open Sphero AI Assistant.

2Type: 'Drive in an equilateral triangle with 1-meter sides. Use blocks only.' Press Send.

3Read the AI's code BEFORE running it. What three headings did it pick? Are they correct for an equilateral triangle? Write the headings in your notebook with your prediction next to them.

An equilateral triangle has 60-degree interior angles, but the EXTERIOR turn (which is what the robot makes) is 120 degrees. Headings: 0, 120, 240. Many AIs get this wrong on the first try.

4Run the AI's code. Did it close the triangle? Measure the gap. If it failed, ask the AI: 'Why did the triangle not close?' Read the answer critically.

5Now ask the AI to FIX its code. Run again. Record both gaps in your notebook: AI v1 gap and AI v2 gap.

An AI that is confidently wrong is more dangerous than one that admits it does not know. Throughout this summer you will hear AIs claim certainty they do not have. Your notebook is the receipt that catches them.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Synthesis - Class Data

Write your team's gap numbers on the class whiteboard table. Now look at the pattern across all pairs.

Copy this table into your Woven notebook BEFORE class data collection starts. Sketch the column headers neatly on a fresh page. As teams report data, fill in your notebook copy AND watch the teacher fill the same table on the whiteboard / slide. Your notebook is your team's permanent record. The projection is shared visibility for the class.

Trial	Code Source	Goal Path	Avg Class Gap (cm)
1	Student	Square	fill in
2	AI v1	Triangle	fill in
3	AI v2	Triangle	fill in

1Which trial had the smallest average gap? Was it always the human, or did the AI sometimes win?

2Calculate the percent change between AI v1 and AI v2 gaps. Did the AI improve when corrected?

Surfaces matter. Code matters. Aim matters. A robot that drives accurately is the product of all three. Engineers call this the 'control loop' - and the rest of this summer is about making your loops tighter.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Robotics Engineer

Day in the Life - Boston Dynamics Robotics Engineer

Watch this video to picture yourself in this career 5 to 10 years from now. What a real robotics engineer's day actually looks like. Same skills you used today, scaled up.

Robotics Engineer: What they do: design and program machines that move through space - on Mars, in warehouses, in surgery suites. Entry pathway: BS in Mechanical or Electrical Engineering or Computer Science with robotics electives. Cal State LA offers a Mechatronics minor and an MS in Robotics. Salary band (BLS, Los Angeles MSA, 2024): entry 78,000 to 95,000. Mid-career 110,000 to 145,000. First step from where you sit today: keep your Sphero programs in a portfolio. Apply to Cal State LA's robotics summer programs.

Spotlight - Dr. Mae Jemison: Dr. Jemison (1956 to present) is an engineer, a medical doctor, and the first Black woman in space (NASA Mission STS-47, 1992). She studied chemical engineering at Stanford, became a Peace Corps doctor in West Africa, and then trained as a NASA astronaut. Her path was multi-discipline before that was a thing. She did not pick engineering OR medicine OR space - she did all three. The robotics work you do today and the medical work you do in Week 4 connect through people like her.

Reflection

1Reflection 1: What surprised you most about programming a robot today?

2Reflection 2: When the AI was wrong, how did you know it was wrong?

3Reflection 3: If you imagine yourself building robots in 8 years, what part of today felt like a glimpse of that future?

I completed Day 1

I have 3 reflection answers in my notebook

I know what 'heading, speed, duration' means

Built by Woven Learning & Technology✎ Edit

Day 2: Logic Gates - If/Then/Else in Motion

Robotics Engineer Pathway - Conditional Logic

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

How Self-Driving Cars See

Watch the video to set up today's thinking - this is the real-world story behind the lab. How autonomous vehicles use logic gates and sensor input to make decisions. Same idea, smaller robot.

Foundations - The Three Words

Yesterday you told the Sphero exactly what to do, step by step. Today you give it a brain. Instead of 'roll forward 2 seconds,' you say 'roll forward UNTIL you bump into something, THEN turn.' That is conditional logic, and it is how every robot, app, and self-driving car decides anything.

The Hook: A Tesla on autopilot sees a traffic cone in the road. In 30 milliseconds it has to decide: brake, swerve, or ignore. The decision is not made by a human. It is made by IF/THEN/ELSE statements running in a chip the size of your thumbnail. Today you write your first one.

IF a condition is true, THEN do this action, ELSE do something different. Example in plain English: IF the floor color is red, THEN stop. ELSE keep rolling. The Sphero BOLT has color sensors on its bottom. It can read the floor in real time.

1Talk to your partner: name three IF/THEN decisions you made this morning before walking in here.

2Predict: if a Sphero is told 'IF speed > 100, THEN turn left,' what speed would NOT trigger a left turn?

I can write an IF/THEN/ELSE statement in plain English

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 Sphero BOLT, iPad with Sphero Edu, red construction paper strips (2), green construction paper strips (2), masking tape, notebook. Build a track on the floor: green = go, red = stop.

Today's task with Sphero Edu: add IF/THEN logic so your Sphero stops when its color sensor sees red - the same conditional logic that runs every autonomous car at every stoplight. Code Here - Sphero Edu: Pick the device in front of you (works on iPhone, iPad, Android, Chromebook, Windows, macOS, or any Chrome browser): - Web app: https://edu.sphero.com/code - All-platforms download page: https://edu.sphero.com/downloads - App stores: search 'Sphero Edu' Sign in: tap 'Join Your Class' and enter the 6-character class code your instructor shares. Connect your robot: hold your BOLT near your device and tap its name when it pops up. The app finds it via Bluetooth.

Basics of Sphero BOLT Coding

Carefully watch this video to learn the tool before you use it. Run-through of Sphero Edu's block canvas, sensor blocks, and IF/THEN statements. Watch this before you build your stoplight robot.

The Hypothesis

1Sketch your track in your notebook. Label where green and red zones are. Predict what code will let your Sphero stop on red and roll on green.

The Build - Stoplight Robot

2In Sphero Edu, drag a Forever Loop block. Inside, add an IF block that asks: 'IF Color Sensor = Red.' Inside the IF, add Stop. Add an ELSE that says Roll forward at speed 50.

3Place your Sphero on green. Press Run. Does it roll? Now slide a red strip into its path. Does it stop?

4Iterate: adjust speed so it does not overshoot the red zone. Find the speed where it always stops on the red strip.

5Challenge: add a second condition. IF green THEN go fast. IF yellow THEN go slow. IF red THEN stop. Test the three-state stoplight.

If your Sphero does not detect the color, lift it, restart the program, and check that the floor is well-lit but not glaring. The sensor is sensitive.

I built a working stoplight with at least 2 conditions

I tested at least 3 different speeds

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Prompt Template - Stoplight Robot: Open Sphero AI Assistant (sparkle icon at the top of your canvas). Prompt template: 'Write a Sphero program that rolls fast (speed 100) on green, slow (speed 40) on yellow, stops on red, and reverses if the tilt is more than 30 degrees. Use a Forever loop.' What to check in the AI's code: - Did it use a Forever loop so the program runs continuously? - Did it nest the IF statements (color check inside the loop, tilt check inside the color check)? - Did it set the right speeds for each color? - Did it forget the ELSE branch?

AI Assistant - Nested Logic

Your stoplight has 2 or 3 conditions. Real robots use dozens. A self-driving car checks color, distance, speed, and angle all at once. Let the AI try.

1Open the Sphero AI Assistant. Type: 'Write a Sphero program that rolls fast on green, slow on yellow, stops on red, and reverses if it tilts more than 30 degrees.'

2Read the AI's code carefully. Did it nest the IF statements correctly? Did it remember to keep the loop going forever? Mark every block in your notebook: correct, wrong, or unsure.

3Run it. Watch what fails. Type back to the AI: 'It did not detect the tilt. Why?' Read the AI's explanation.

4Ask the AI to fix it. Run again. How many tries did it take to get a working program?

Real engineers use AI to draft code, then audit every line. They never deploy AI code without testing. You just did the same thing - that is the actual workflow.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Synthesis - Prompt Quality vs Iteration Count

Pair	Prompt Length	AI Iterations Needed	Final Status
Yours	fill in	fill in	works / partial / fails
Class avg	share	share	share

1Write the EXACT prompt that finally worked. What did it have that your first prompt did not?

Engineers call this 'prompt engineering.' The clearer your specification, the fewer the iterations. This is a real job skill.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Robotics Engineer - Conditional Logic

Day at Work: Robotics Engineer

Watch this video to picture yourself in this career 5 to 10 years from now. What a robotics engineer's actual day looks like - the conditional logic you wrote today is core to her job.

Robotics Engineer (Software side): What they do: write the decision logic that lets robots react to the world - color, sound, distance, force. Entry pathway: BS in Computer Science or Computer Engineering with control-systems coursework. Cal State LA's BS CS + Robotics minor or MS in Mechatronics covers this. Salary band (BLS, Los Angeles MSA, 2024): entry 82,000 to 98,000. Mid-career 115,000 to 150,000. First step from where you sit today: build a portfolio on GitHub starting with a Sphero project this week.

Reflection

1Reflection 1: When did your code do something you did not expect? What did you learn?

2Reflection 2: What is one IF/THEN decision you make every day without thinking?

3Reflection 3: How would you describe the difference between human logic and code logic to a 4th grader?

I completed Day 2

I built a multi-conditional program

I have 3 reflection answers

Built by Woven Learning & Technology✎ Edit

Day 3: Secret Missions - Pathfinding Algorithms

Robotics Engineer Pathway - Algorithmic Thinking

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

A* Pathfinding Visualized

Watch the video to set up today's thinking - this is the real-world story behind the lab. A clean visualization of how a search algorithm finds the shortest path. We are going to do this with code today.

Foundations - Three Pathfinding Strategies

Yesterday you taught your robot to make decisions. Today you teach it to PLAN. Here is what we are doing today: you and your partner will BUILD A MAZE on the floor with tape, boxes, and string. Then you will program your Sphero to find its way from Start to Goal three different ways. The fastest team gets bragging rights. Why a maze? Because every robot, every video game character, every Google Maps route uses the same kind of pathfinding code you are about to write.

The Hook: Google Maps checks 7 billion possible routes between any two points in the world before it shows you one. It does this in under 0.4 seconds. The algorithm it uses is over 60 years old - invented by a mathematician named Edsger Dijkstra who scribbled it on a napkin in a coffee shop in 1956. Today's mission: you BUILD a maze on the classroom floor (tape, boxes, string), then code your Sphero to drive it three ways - shortest path, fastest path, and AI's path. The team that beats their own best time wins.

Strategy 1: Greedy. Always head straight at the goal. Fast but bumps into walls. Strategy 2: BFS (Breadth-First Search). Try every path level by level. Slow but always finds something. Strategy 3: Dijkstra. Try paths in order of total cost. Fastest known way to find the GUARANTEED shortest path. Google Maps uses a flavor of Dijkstra. So does your video game's enemy AI. So does the routing on a UPS truck.

1Talk to your partner: when you walk from your front door to your favorite spot in your neighborhood, do you take the shortest path or the easiest path? Are they the same?

2Predict: in your maze, will the SHORTEST path always be the FASTEST path? Why or why not?

I can name 3 pathfinding strategies

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 Sphero BOLT, iPad, masking tape, 4-6 cardboard boxes as obstacles, plus string or red yarn (for laser-grid effect), 3 small objects (the 'targets' to retrieve), ruler, notebook. Floor maze: 2 m by 2 m square with 4 obstacles inside. Start at one corner, Goal at the opposite corner.

Today's task with Sphero Edu: write the code that solves a physical maze - your first real pathfinding algorithm. Code Here - Sphero Edu: Pick the device in front of you (works on iPhone, iPad, Android, Chromebook, Windows, macOS, or any Chrome browser): - Web app: https://edu.sphero.com/code - All-platforms download page: https://edu.sphero.com/downloads - App stores: search 'Sphero Edu' Sign in: tap 'Join Your Class' and enter the 6-character class code your instructor shares. Connect your robot: hold your BOLT near your device and tap its name when it pops up. The app finds it via Bluetooth.

Mission Requirements - Your Maze MUST Have: Build: - At least 3 obstacles (boxes, books, cardboard towers) - A laser maze section (string or red yarn stretched at Sphero height to dodge) - At least one secret passage (a hidden route the Sphero can take) Code: - Minimum of 10 blocks total - At least one DIRECTION block (Roll) - At least one LOOP block (Repeat or Forever) - At least one COLOR block (Set Main LED, or color-sensor IF) - At least one SENSOR block (Color Sensor, IR, or Tilt) - A 'secret coded message' delivered via the LED matrix or color sequence Tip: use the 8x8 LED matrix on the BOLT to flash a color code. Make a key (red = stop, green = clue 1, blue = clue 2). The color sequence IS your secret message.

The Hypothesis

1In your notebook, sketch the maze top-down. Mark Start, Goal, and obstacles. Sketch THREE possible paths from Start to Goal. Then PREDICT: which of your 3 paths will solve the maze fastest? Circle that path, and write a one-sentence reason (shortest? fewest turns? safest from obstacles?).

The Build - The Secret Mission Maze

How to Use the Sphero Pathfinder Simulator (below): This simulator is your VIRTUAL MAZE TRAINER. Use it BEFORE you spend 10 minutes coding your real Sphero through the physical maze you built. How: 1. Pick one of the 3 preset challenges (Easy / Medium / Hard) at the top. 2. Drag Forward, Turn Left, and Turn Right blocks into the sequencer panel on the right. 3. Click 'Run' to watch the simulator's Sphero attempt your path. 4. If the Sphero crashes into a wall, the screen tells you which step failed. Edit and re-run. 5. Once you can solve a preset cleanly, you have the pathfinding intuition you need to code your real Sphero through the physical maze. Treat this as a 5-minute practice round, then go to your physical maze.

2Open the Sphero Pathfinder simulator (above). Drag obstacles to match your physical maze. Run the simulator's pathfinding to see ONE possible solution path before you write your own code - this is your reference, not your answer.

3Now write your OWN pathfinding code in Sphero Edu. Compare: did you take the same path as the simulator? A faster one? A safer one? Record your route and timing in your notebook.

4Predict: which of your 3 paths is the SHORTEST in distance? Which is the FASTEST in time given the Sphero needs to slow down to turn?

5Pick your shortest-distance path. Code it as a sequence of Roll blocks. Time it with a stopwatch from start to goal.

6Now code your fastest-TIME path - usually the one with fewer turns, even if it is longer. Time it.

7Record both times in your notebook. Calculate the difference.

8Trade mazes with another team. Solve theirs in under 90 seconds of code time. Run their solution. Did it work?

9When you finish testing your maze, swap with another team. Run their maze. Tell them ONE thing that worked and ONE thing that could be better. Listen to their feedback on yours. This is the most important part.

Watch the corners. Sphero turns are imperfect at high speed. A faster speed often means a wider turn radius and you smash the box.

I tested at least 2 different paths through my maze

I recorded times for each

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI as Pathfinder

1Describe your maze to the Sphero AI Assistant in plain English. Include obstacle positions in coordinates: 'Obstacle at (50 cm, 50 cm), size 20 cm by 20 cm. Start at (0, 0). Goal at (200 cm, 200 cm).'

2Ask: 'Generate the shortest Sphero path that avoids all obstacles.'

3Read the path. Trace it in your notebook on top of your maze sketch. Did the AI's path actually avoid the obstacles, or did it cut through one?

4Run the AI's code. Time it. Compare to your human-built fastest-time path. Who won?

AI is bad at spatial reasoning right now. Even Gemini and Claude routinely get coordinate-based mazes wrong. This is one of the open problems in AI research today. You just hit a frontier.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Synthesis - Human vs AI Pathfinding

Solver	Path Distance (cm)	Time to Goal (sec)	Crashes
Human v1	fill	fill	fill
Human v2	fill	fill	fill
AI	fill	fill	fill

1Which solver had the cleanest run? Which had the shortest path? Were they the same solver?

In real robotics, humans plan the high-level strategy and AI executes the low-level path tweaks. The combination beats either alone. Today you saw why.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Autonomous Systems Engineer

Careers in Robotics Engineering

Watch this video to picture yourself in this career 5 to 10 years from now. A working engineer building autonomous robots for a recycling plant. Pathfinding code at industrial scale.

Autonomous Systems Engineer: What they do: write the algorithms that let warehouse robots, drones, and self-driving cars find their way through a real-world space. Entry pathway: BS Computer Science or Computer Engineering with algorithms and AI coursework. Cal State LA offers AI-focused electives in the CS department. Salary band (BLS / Glassdoor LA 2024): entry 90,000 to 110,000. Mid-career 130,000 to 175,000. Senior 200,000 plus. First step from where you sit today: solve all 30 problems on LeetCode's BFS/DFS section by end of summer. Free.

Reflection

1Reflection 1: What surprised you about how the AI found (or failed to find) a path?

2Reflection 2: When have you taken a longer route on purpose? Why?

3Reflection 3: If algorithms decide your route, your search results, and your news feed - how do you know if they are picking well for you?

Spotlight: Dr. Gladys West: Dr. West (1930 to 2026) was a mathematician whose work is inside almost every robot today. She did the complex math and programming that created the foundation for GPS - the Global Positioning System. When you program a Sphero to move to a specific coordinate or follow a path, you are using technology that Dr. West helped invent. Pathfinding algorithms exist on top of her math. For most of her career, her name was barely known. She was inducted into the Air Force Hall of Fame at age 88. Brilliance is sometimes invisible until someone bothers to look.

4Reflection 4 (peer credit): name ONE thing you saw another team do better than you. Name ONE person who helped you today. Thank them out loud before you leave.

I completed Day 3

I beat the AI on at least one path

Reflection answers in notebook

Built by Woven Learning & Technology✎ Edit

Day 4: Speedway Showdown - Human vs AI

Robotics Engineer Pathway - Human-AI Collaboration

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Today's goal: race ONE classroom maze three different ways. First with only your own code. Then with only AI's code. Then with you and the AI working together. Best time wins.

The Hook: In 2017 the world's best Go player Lee Sedol lost to an AI called AlphaGo. The world thought humans were done. Two years later, a chess study found that a mid-skill human PLUS a mid-quality AI beat both top humans alone AND top AIs alone. The lesson: collaboration wins. Today you test that yourself with a Sphero, a maze, and three different ways of writing the code that solves it.

How Self-Driving Cars Decide

Watch the video to set up today's thinking - how autonomous systems balance human input with AI suggestions in real-time. Today we test the same tradeoff with our Spheros.

Foundations - The Three Modes

Mode 1 - Pure Human: you write every line of code yourself. AI Assistant turned off. Mode 2 - Pure AI: you describe the goal in one prompt and run whatever the AI gives you, without edits. Mode 3 - Collaboration: you generate with the AI, then audit, edit, and iterate before running. Same maze. Three modes. Best time wins. Crashes count as DNF (did not finish).

1Predict: which mode do you think will be fastest? Why?

I understand the three modes of today's challenge

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 Sphero BOLT, 1 iPad with Sphero Edu, stopwatch, notebook. ONE shared maze at the front of the room (your instructor sets it up - same maze for every team, so times are comparable). Whiteboard scoreboard at the front with 3 columns: Human, AI, Collab.

Today's task with Sphero Edu: write your fastest pathfinder code, then race it head-to-head against an AI-written version (you bring AI in during Phase 3). Code Here - Sphero Edu: Pick the device in front of you (works on iPhone, iPad, Android, Chromebook, Windows, macOS, or any Chrome browser): - Web app: https://edu.sphero.com/code - All-platforms download page: https://edu.sphero.com/downloads - App stores: search 'Sphero Edu' Sign in: tap 'Join Your Class' and enter the 6-character class code your instructor shares. Connect your robot: hold your BOLT near your device and tap its name when it pops up. The app finds it via Bluetooth.

The Hypothesis

1Predict your time for each mode. Write 3 numbers in your notebook before you start.

Set up your team's SHOWDOWN SCOREBOARD now. Turn to a fresh page in your Woven notebook. Draw a 4-column table with these headers: Mode | Predicted time | Actual time | Notes. Add 3 rows labeled Pure Human, Pure AI, Collaboration. You will fill this in as you run each mode below. The class also tracks combined times on a whiteboard up front - but your notebook scoreboard is your team's permanent record. (The embedded Showdown app's scoreboard cannot be filled in or saved - your notebook is the real one.)

Mode 1 - Pure Human

2Close the AI Assistant. Code the maze solution from scratch. Time from 'start coding' to 'Sphero crosses goal line'. Write the time in your notebook SHOWDOWN SCOREBOARD (Pure Human row, Actual time column) AND on the class whiteboard so other teams can see.

Mode 2 - Pure AI

3Reset. Open AI Assistant. Write ONE prompt describing the maze and the goal. Run whatever it generates without editing. Time it. If it crashes, that is real data - record DNF. Write the time (or DNF) in your notebook SHOWDOWN SCOREBOARD (Pure AI row) AND on the class whiteboard.

Mode 3 - Collaboration

4Reset. Use AI Assistant to draft. Read the code. Edit anything that looks wrong. Iterate up to 3 times. Run. Time it. Write the time in your notebook SHOWDOWN SCOREBOARD (Collaboration row) AND on the class whiteboard.

5Compare your 3 times to your predictions. Which mode actually won?

DNFs count. If pure AI fails, that is real data. Half of pure-AI runs in real software engineering also fail.

I ran all 3 modes

All 3 times on the scoreboard

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

Why this app: today in Phase 2 you wrote your OWN maze code. Now run the same maze 3 ways back-to-back - Pure Human, Pure AI, Collaboration - and see whose pathfinder wins. The Sphero Pathfinder simulator (below) runs each 60-second match for you and posts times to a class scoreboard. You are not learning the simulator; you are using it as a stopwatch + arena to test the question: who is faster, you or the AI? Run all 3 modes before reading the steps below.

1Open the Showdown app (above). Run all 3 modes back-to-back: Pure Human (your code only), Pure AI (AI code only), Collaboration (you write, AI tweaks). After each run, the app shows your time. Since the app's scoreboard cannot be saved, copy each time into your notebook SHOWDOWN SCOREBOARD (the page you set up in Phase 2) AND on the class whiteboard.

2Compare your team's 3 times. Which mode was fastest? Which was most consistent? In your notebook, write: 'In a real life-or-death situation - search and rescue, autonomous medical delivery - I would trust ___ because ___.'

What Just Happened

3Look at the class whiteboard scoreboard - all teams' times, all 3 modes. How many pairs had Collaboration win? Pure Human win? Pure AI win? Add a tally line to your notebook scoreboard underneath: 'Class results: Collab won __ / Human won __ / AI won __.'

4Find a pair where Pure AI won. Ask them: what was special about their prompt?

5Find a pair where Pure Human won. Ask: did they have prior coding experience?

The most interesting result is not who won. It is the SPREAD between Pure AI's best and worst times. AI is high variance. Humans are lower variance. Collaboration combines high ceiling with stability.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Week 1 Synthesis

Mode	Class Best Time (sec)	Class Worst Time (sec)	DNF Count
Pure Human	fill	fill	fill
Pure AI	fill	fill	fill
Collab	fill	fill	fill

1Calculate the spread (worst minus best) for each mode. Which mode had the smallest spread? That is the most reliable mode.

In professional software engineering, this exact tradeoff drives team decisions: Pure AI is fast when it works but unreliable. Humans are reliable but slow. The mix is where careers live.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Robotics Engineer - Human-AI Collaboration

What is it Like to Work as a Robotics Engineer

Watch this video to picture yourself in this career 5 to 10 years from now. Real-world advice from someone who does this job day to day. Same human-AI tradeoffs you tested today.

ML / Robotics Engineer (Human-in-the-Loop Systems): What they do: design systems where humans and AI work together - approval workflows, AI tools that need expert audit, prompt engineering teams. Entry pathway: BS in CS, Data Science, or Cognitive Science. Cal State LA's BS in CS with the AI emphasis fits exactly. Salary band (Glassdoor LA 2024): entry 95,000 to 115,000. Mid-career 140,000 to 180,000. First step from where you sit today: read 'Human + AI = Better Decisions' free articles on MIT Sloan Review.

Reflection

1Reflection 1: Did your prediction about which mode would win match the result? Why or why not?

2Reflection 2: When in your life is collaboration faster than working alone?

3Reflection 3: Looking back at this week, what is one thing you can now do that you could not do on Monday?

Next week we trade 1-pound robots for 30-gram drones. The physics is different. The stakes are higher. Bring your notebooks - everything you learned this week applies.

Week 1 Complete: You spent four days teaching a robot to think with logic, conditionals, and pathfinding. You also went head-to-head with an AI and either won, lost, or proved that collaboration beats both. Stand up. Find your partner. Take a group photo. You finished Week 1. Group photo. Sign your notebook. See you Monday for Week 2 - drones.

I completed Week 1

Reflection answers in notebook

I am ready for drones

Built by Woven Learning & Technology✎ Edit

Day 5: First Flight - APEX 149 Pilot Training

Aeronautical Engineer Pathway - Manual Flight + Pilot Skills

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

APEX 149: Drone Unboxing and Setup

Watch this short video to see why today's topic matters before we dive in. Woven Learning's own walkthrough of the APEX. Watch start to finish before you touch the drone.

Foundations - The Four Stick Inputs

Welcome back.

The Hook: In 1903 the Wright brothers flew 120 feet at Kitty Hawk in 12 seconds. They had spent 4 years studying birds and failing wind-tunnel tests. Their breakthrough was not lift - it was CONTROL. They built the first 3-axis control system: pitch, roll, yaw. Every drone you fly this summer uses those exact same 3 axes - plus a fourth: throttle. 122 years later, the framework has not changed. Today you become the pilot.

Your APEX controller has TWO joysticks. Memorize this map: LEFT STICK - Up/Down = THROTTLE (altitude). Push up to climb, push down to descend. - Left/Right = YAW (rotation). Push left to spin counter-clockwise, right to spin clockwise. RIGHT STICK - Up/Down = PITCH (forward/backward tilt). Push up to fly forward, down to fly backward. - Left/Right = ROLL (side-to-side slide). Push left to slide left, right to slide right. Small, gentle movements. Always.

1Talk to your partner: when an airplane banks to turn, which axis is it using? When a helicopter spins in place, which axis?

2Predict: if the throttle is at 50 percent and you let go of every other stick, what does the drone do?

I can name the four stick inputs (throttle, yaw, pitch, roll)

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Safety Zone Briefing - READ ALOUD AS A GROUP: Clear Communication: shout 'Drone in course!' before you launch. Shout 'Pilot ready!' before you take a turn. One At A Time: only one drone flies the course at a time. Eyes On The Drone: always keep your drone in sight when flying. Crash Plan: if you lose control or are about to crash, push the LEFT STICK ALL THE WAY DOWN immediately. That cuts throttle and grounds the drone safely. No-Fly Zones: students waiting their turn sit on a designated bench. Drones never fly over the bench.

Materials per pair: 1 APEX 149 drone, 1 controller (with 2 fresh AAA batteries), 1 fully-charged drone battery, 1 smartphone with the Tspeed 7 app installed (only needed Day 12 onward), 1 set of propellers (4 props), eye protection, taped 3 m by 3 m flight zone, 3 hula hoops (or large rings), notebook.

Today's task with the APEX 149: connect the drone, pair the physical controller, and run your first hover - no code yet, just stick skills, like a real pilot's first flight. Code Here - APEX 149 Pilot Apps: Today you fly the APEX with the PHYSICAL CONTROLLER. You will pair the drone to the controller and use the joysticks. The phone app is NOT used today (the manual says drone is paired to ONE thing at a time - controller OR app, never both). On Day 12 you switch to the app for the FPV mission. For now, set the phone aside.

App for later (Day 12 onward): - iOS: Tspeed 7 - https://apps.apple.com/in/app/tspeed-7/id1575505851 - Android: https://play.google.com/store/apps/details?id=com.apex.at149 - Chromebook / Windows: AT-66BL from Microsoft Store / Google Play - Or browser: https://echo.pitsco.com (Pitsco's web-based block coding for the same drone)

APEX 149 Coding Drone Tutorial - Setup and Propellers

Carefully watch this tutorial to learn the steps before you try them. APEX's official walkthrough. Watch the propeller install section closely - we install ours next.

Pre-Flight Checklist (Read Aloud)

Pre-Flight Inspection - DO THIS ON YOUR ACTUAL DRONE before every flight today (and every day this week). One partner reads each item out loud, the other partner physically checks it on the drone. Switch roles for the next inspection. Propellers (4 total - check each): - Spin freely, no chips or cracks on the leading edge - Prop nut snug (cannot wiggle the prop with a fingertip) - Mounted in the correct rotation (CW prop on CW position, CCW on CCW) Motors (4 total - spin each one with a finger, drone OFF): - Spins free and quiet, no grit or scrape - No lint, hair, or carpet fiber wrapped on the shaft Battery: - Flat (NOT puffy or bulging - if puffy, GROUND it and put in LiPo bag) - 3 of 3 green LED bars (full charge) - Velcro strap tight, battery does not shift when you wiggle the drone Frame + body: - No cracks on any of the 4 arms (especially at the joints) - All motor mount screws flush and tight - Antenna straight up, not bent or kinked - Camera lens clean (microfiber cloth ONLY, never your shirt) Controller: - 2 fresh AAA batteries seated correctly - Power LED solid green - Beep is clear and strong on power-up (weak/raspy = swap batteries) Bind check (last): - Drone status LED is FLASHING blue (means bound to controller) - If solid blue: re-bind by holding both joysticks to lower-right corner for 3 seconds Any GROUND-level item = do not fly. Any NOTE-level item = log it in your notebook and tell the instructor.

Step 1 - Pair Drone + Controller

1Insert 2 AAA batteries in the back of the controller. Slide the drone's battery onto the drone body. Place the drone on a flat surface.

2Press and hold the power button on the controller until it beeps and the power button light comes on.

3Press and hold the power button on the drone battery. Three green battery-level LEDs come on (full charge).

4PAIR: move BOTH joysticks to the LOWER-RIGHT corner at the same time and hold. The controller will beep. The blue LED on top of the drone goes from rapid flash to slow flash. Pairing is now done (one-time).

Step 2 - Arm + Take Off

5ARM: move LEFT joystick to LOWER-LEFT corner and RIGHT joystick to LOWER-RIGHT corner at the SAME TIME. Hold for 2 seconds. The motors begin to spin slowly. This is your armed state.

6TAKE OFF: release the joysticks and IMMEDIATELY press the Take Off button on the controller (within 2 seconds, while motors are still spinning slowly). The drone rises to about 1 meter and hovers.

7LAND: press the Take Off / Land button again to land automatically. The drone descends gently to the ground and shuts off all motors.

Step 3 - First Hover Drill

Hypothesis: how much will your APEX drift from its takeoff spot during a 10-second hover with no input on pitch or roll? Predict a number in cm in your notebook before flying.

8After takeoff, leave the joysticks in the middle for 10 seconds. The drone holds altitude (this is HOVER STATE). Measure how far it drifted from start. Record.

9Practice altitude: push LEFT joystick UP slowly (drone climbs) - release - drone holds new height. Pull DOWN slowly - drone descends - release - drone holds. Repeat 3 times.

10Practice yaw: push LEFT joystick LEFT slowly (drone spins counter-clockwise). Push RIGHT (clockwise). Re-center the drone facing forward.

11Practice forward/back: push RIGHT joystick UP (drone glides forward). Pull back (drone glides backward).

12Practice roll: push RIGHT joystick LEFT and RIGHT (drone slides side to side). Land.

Step 4 - Hoop Pass Drill

13Partner holds a hula hoop vertically at chest height. Take off, fly forward through the hoop, hover on the other side, return through the hoop, land. Two passes.

Trim Buttons - Use Them If The Drone Drifts: If the drone keeps shifting forward when you are not pushing the stick, press the BACKWARD trim arrow on the controller. If it drifts left, press RIGHT trim. The trim adjusts the drone's resting balance until you stop touching it. This is an actual flight-engineer skill. Real drones drift. Trim is how you fix it.

Crash protocol: if your drone is about to hit a wall or person, LEFT STICK ALL THE WAY DOWN immediately. Throttle = zero. Drone falls and powers off safely. Practice this reflex BEFORE you need it.

I paired controller + drone correctly

I armed and took off using the official sequence

I hovered for 10 seconds and recorded drift

I flew through a hula hoop forward and back

I used the trim buttons at least once

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

Why an AI Cannot Fly Your Drone Today

AI Prompt Template - Drone Physics: Copy and paste this exact prompt into Gemini (https://gemini.google.com) or Claude (https://claude.ai): 'Explain quadcopter pitch, roll, and yaw to a high school student. Use a real-world example for each.' Then ask follow-up: 'Why do quadcopters need 4 motors instead of 3?'

1Open Gemini or Claude. Ask: 'Can you remotely control my APEX 149 drone using my phone?' Read the answer.

2Now ask: 'What sensors and controls would an AI need access to in order to fly a drone autonomously?' Note the answer (camera, IMU, GPS, motor outputs, control loop).

3Ask: 'Why can a Tesla self-drive but my hobby drone cannot?' Listen for: SAFETY CERTIFICATION, REDUNDANT SENSORS, REGULATION, LIABILITY.

Today YOU were the controller. Next week you will write code that flies the drone for you. By Day 13, perception (camera) will trigger action (movement) in a closed loop. Step by step, you become a pilot. Then you become an autonomy engineer.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Hover Drift Data

Hover Drift Across the Class (cm over 10 seconds)

Pair	Predicted drift	Actual drift	Difference
A
B
C
D

1What is the average drift? Which pair was closest to predicting reality?

Real drone engineers call this 'station-keeping' performance. A consumer drone hovers within 30 cm. A military drone hovers within 5 cm. You just measured your APEX. That number is your baseline for every test this week.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Aeronautical Engineer + Drone Pilot

9 Growing Jobs for Drone Pilots in 2025

Check out this video to see a real professional in this role. From the Drone Nerds channel. Real entry-level paths people are taking right now.

Aeronautical Engineer / Commercial Drone Pilot: What they do: design, test, and operate aircraft. With an FAA Part 107 certification (a written test you can take at age 16), you can be a paid commercial drone pilot. Entry pathway: BS Aerospace or Mechanical Engineering for the engineer track (Cal State LA + Cal Poly Pomona). For the pilot track: 14-week prep + Part 107 test, then real-world hours flying for surveying, real estate, film, search-and-rescue. Salary band (BLS LA 2024): aerospace engineer entry 84,000 to 102,000. Commercial drone pilot entry 50,000 to 75,000 (often hourly, 50 to 150 per hour for specialty work). First step from where you sit today: study for the Part 107 written test. Practice tests are free online.

Spotlight: Ernest Levert: Levert (1954 to present) is an aerospace engineer at Lockheed Martin who pioneered ROBOTIC WELDING for the International Space Station and NASA's Space Shuttles. He proved robots could do precision construction at scale - exactly the kind of automation that lets your APEX exist as an affordable kit today. His path: rural Mississippi to Tennessee Tech to Lockheed. Underrepresented in his industry, he led teams that built billion-dollar systems. The robot you flew today exists because his generation made the manufacturing possible.

Reflection

1Reflection 1: what surprised you about flying your APEX?

2Reflection 2: when have you trusted a checklist before doing something risky?

3Reflection 3: tomorrow we add weight. What do you think will happen?

I completed Day 5

I logged my drift number

My drone is back in the charging dock

Built by Woven Learning & Technology✎ Edit

Day 6: The Weight-to-Lift Inquiry

Aeronautical Engineer Pathway - Manual Flight Testing

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

How Does A Wing Actually Work? (Veritasium)

Carefully view this video to ground today's hook in a real example. Veritasium digs into the actual physics of how wings generate lift - then debunks the simplified explanation you may have heard in textbooks. Watch the full video.

Foundations - The Equation

Yesterday you connected the drone, paired the controller, and ran your first hover. Today you measure how much weight it can carry. Today you fly. Before you do, you ask the question that drives every aeronautical engineering team on earth: how much can it lift, and how heavy is it? That ratio decides everything.

The Hook: A Boeing 747 weighs 412,000 pounds empty. Fully loaded with fuel, cargo, and passengers it weighs 875,000 pounds. To lift it, the four engines produce 252,000 pounds of THRUST - and the wings convert it to over 1 MILLION pounds of LIFT. Lift-to-weight ratio: 1.14. Just enough to fly. Your APEX 149 weighs around 100 grams. With max throttle, the four motors produce enough lift to climb fast even with extra weight. Today you measure exactly how much weight your APEX can carry before climb performance breaks down.

Lift = total upward force from spinning propellers (measured in grams or Newtons). Weight = mass times gravity (just the mass in grams works for our purpose). Lift-to-Weight Ratio = Lift / Weight. If ratio < 1: drone cannot leave the ground. If ratio = 1: drone hovers, cannot climb. If ratio > 1: drone climbs. Higher ratio = faster climb.

1Predict: what happens to your APEX 149's lift-to-weight ratio if you tape a 5-gram weight to it?

I understand the lift-to-weight equation

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Safety Refresher: eyewear ON. Stay in your taped flight zone. If anything goes wrong - LEFT STICK ALL THE WAY DOWN. Battery low warning means LAND IMMEDIATELY.

Materials per pair: APEX 149 + controller + phone with Tspeed 7, fresh fully-charged battery, kitchen scale (0.1 g resolution), 5 paper clips (each weighs about 1 gram), small piece of painters tape, stopwatch, eyewear, taped 3 m by 3 m flight zone, notebook.

Drone Setup - Same as Day 5: Pair the controller with the drone (both joysticks to lower-right corner, hold). Phone app NOT needed today (still using physical controller). If you forgot the pairing sequence: see Day 5's notebook page or ask your instructor for a quick refresh.

The Question

How does adding weight affect your drone's ability to climb? Today you make TWO predictions of your drone's lift threshold - your own gut, and an engineer's calculator - then run real flight trials to see who was closest.

1Weigh your drone (no payload) on the kitchen scale. Record the empty mass in grams. (Most APEX 149s weigh between 28-32g.)

Your Hypothesis - Your Gut Prediction

2In your notebook, write your GUT prediction: at what payload (in paper clips, 1g each) do you think your drone will fail to climb at full throttle? Hint: light drones typically lift 25-50% of their own weight. Yours weighs ~30g. Write a number + one sentence of reasoning.

The Engineer's Tool - Drone Lift Calculator

Why this app: real engineers do not just guess - they MODEL. The Drone Lift Calculator (below) takes your drone's mass, payload, motor count, and thrust per motor, and predicts whether the drone will HOVER, BARELY HOVER, or be GROUNDED. Use it to make your second prediction (the calculator's prediction) before you fly. After your real trials, you will compare gut vs calculator vs reality.

3Open the calculator (above). Enter your drone's empty mass from your weigh-in. Set Payload mass to 0g. Note the lift-to-weight ratio - this is your no-payload baseline.

4Now slide Payload mass up in 1g steps: 1g, 2g, 3g, 4g, 5g. At what payload does the status flip from HOVER ✅ to BARELY HOVER? At what payload does it flip to GROUNDED ❌? Record both numbers in your notebook as the CALCULATOR's PREDICTION.

5Compare your two predictions: how close was your GUT prediction to the CALCULATOR's prediction? Off by 1g? 2g? Same number? Write the comparison in your notebook before you fly.

The Build - Real Climb Test

6Predict: how high (in body lengths) can your drone climb in 5 seconds at full throttle with NO payload?

7Trial 0 (baseline, no payload): on the controller, push the LEFT STICK to FULL throttle for 5 seconds. Watch how high the drone climbs. Estimate climb height in body lengths (1 body length is roughly 12 cm).

8Tape 1 paper clip (about 1 gram) to the top of the drone, dead center. Reweigh. Repeat the climb test. Record.

9Repeat with 2 paperclips. Then 3. Then 4. Each trial: weigh first, fly second, record climb height.

10Compare ALL THREE: at what REAL payload did your drone fail to climb? How close was your GUT prediction? How close was the CALCULATOR's prediction? If they all differ, write a 1-sentence hypothesis - what physics is the model missing? (Battery sag? Motor wear? Prop angle?)

11Build a data table: Trial, Total Mass (g), Climb Height (body lengths in 5 sec), Notes.

If the drone cannot leave the ground with the added weight, that is real data. Record it as 'failed to lift' and remove a clip. The line where lift fails is your drone's ceiling.

I ran 5 climb trials at increasing mass

I have a complete data table

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Predicts the Curve

1Open Gemini. Type: 'I have a small quadcopter that weighs about 100 grams empty. With each 1-gram paperclip I add as payload, what general trend should I expect for climb height in 5 seconds at full throttle?'

2Write the AI's predicted trend in your notebook (linear? curved? at what mass does it predict failure?).

3Compare to your actual data. Was the AI's general trend right? Was its failure-point prediction right?

4Ask: 'What real-world variables make actual climb performance worse than a model predicts?' Note: battery age, motor wear, propeller damage, room air currents.

The AI is great at directional reasoning ('more weight equals less climb') but bad at exact numbers without your drone's spec sheet. Real engineers always test.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Lift Curve

Class Climb Performance vs Payload

Pair	Empty mass	1g	2g	3g	Failure point
A
B
C
D

1Sketch a graph: payload mass on X, climb height on Y. Draw your data and the class average. What is the shape?

Engineers call this the THRUST CURVE. Every aircraft has one. Cargo plane operators use it to decide max payload. You just measured yours.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Flight Test Engineer

What Does an Aerospace Engineer Do?

Watch this video to see what a real day in this career actually looks like. A NASA aerospace engineer walks through her day - rockets, airplanes, simulations, lab work.

Flight Test Engineer: What they do: design and run controlled experiments on aircraft and drones to verify they fly the way the math says. Entry pathway: BS Aerospace or Mechanical Engineering. Cal State LA + Cal Poly transfer track is the local route. Salary band (BLS LA 2024): entry 88,000 to 105,000. Mid-career 125,000 to 165,000. First step from where you sit today: AIAA student chapter at Cal State LA, plus DIY drone projects you can document in a portfolio.

Spotlight: Lonnie Johnson: Johnson (1949 to present) is famous for inventing the Super Soaker, but he is also a NASA aerospace engineer. He spent his career designing power systems for the Galileo mission to Jupiter and the Cassini mission to Saturn - keeping robotic explorers running for years in deep space. In high school, he built a remote-controlled robot named Linex from junkyard scraps. The same DIY spirit you used to tape a paperclip to your drone is what got him to NASA.

Reflection

1Reflection 1: what did watching your drone struggle to lift feel like vs what the data showed?

2Reflection 2: when did your prediction match reality? When did it not?

3Reflection 3: a 1-gram paper clip changed your drone. What other tiny changes have huge effects in the world?

I flew my first weight-to-lift trial

I have a thrust curve sketched

Built by Woven Learning & Technology✎ Edit

Day 7: Payload Variable Testing

Aeronautical Engineer Pathway - Controlled Variable Experiments

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

The Physics of Drones

Watch the video to set up today's thinking - this is the real-world story behind the lab. How payload affects flight. We will measure this on our drones today.

Foundations - Variables

Yesterday you measured how mass affects climb. Today you ask the harder question: what about WHERE you put the mass? A drone with the weight on the front behaves nothing like one with weight on the back.

The Hook: A cargo plane carrying 100,000 pounds of cargo cannot just stack it anywhere. If the load shifts to the back, the plane noses up and stalls. If it shifts forward, it dives. There is a tiny zone called the CENTER OF GRAVITY where the plane is balanced. Move out of it by 6 inches and the plane crashes. Your APEX 149 has the same center of gravity rule, scaled down.

INDEPENDENT variable: the one thing you change. DEPENDENT variable: what you measure as a result. CONTROLLED variables: everything you keep the same. Today's experiment: Independent: position of payload (front, back, center). Dependent: hover stability (drift in cm over 10 seconds). Controlled: payload mass, battery state, takeoff height, room conditions.

1Predict: where will weight cause the drone to drift the most? Front, back, or sides?

I can name the 3 types of variables

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: APEX 149 + controller + phone with Tspeed 7, fresh battery, kitchen scale, 1 paper clip (2 g), painters tape, ruler, taped 3 m by 3 m flight zone with cm grid drawn on paper at center, eyewear, notebook. Fresh batteries at the charging station.

Today's task with Pitsco Echo: add measured payload weights to the drone, then test how each weight changes hover stability and drift. Same Echo app you will fly autonomously starting Day 10 - today you fly manually with payload. Code Here - Pitsco Echo (block coding for APEX): Web app: https://echo.pitsco.com (Chrome on a laptop or tablet). Sign in with the class code your instructor shares. Open Tspeed 7 on your phone first, then connect Echo to the drone via the phone's Bluetooth. Drag blocks from the left palette into the canvas. Click RUN. The drone flies your code. The OFF switch on the drone or controller is your kill button.

APEX 149 Coding Drone Tutorial (First Time Using Echo)

Carefully watch this tutorial to learn the Pitsco Echo block-coding interface BEFORE you build your first program. This is the same Echo you will use Days 7-11 - learn the layout (block palette, canvas, run button, drone connection) once and you're set for the week.

The Hypothesis

1Sketch the drone top-down with X/Y axes. Mark 4 payload positions: Front (+Y), Back (-Y), Left (-X), Right (+X). Predict drift direction for each.

The Build - Hover Stability Test

2Open the calculator (above). Enter your drone's mass + 2g payload. Confirm the status reads HOVER ✅. This proves your drone CAN lift today's payload - so any drift you see in the trials below is from POSITION, not from being underpowered.

3Tape paper clip on FRONT of drone. Run program: Takeoff (0.8 m) → Wait 10 s → Land. Watch where the drone drifts. After landing, measure how far from center it ended in cm. Record.

4Move clip to BACK. Repeat. Record drift distance and direction.

5Move clip to LEFT side. Repeat.

6Move clip to RIGHT side. Repeat.

7Bonus: tape clip dead CENTER (on top of flight controller). Repeat. This is your 'control' - drift should be smallest.

8With 4 drift readings in hand, look back at the calculator's HOVER ✅ confirmation from earlier. Lift was never the problem - so 100% of today's drift is from offset weight (torque). The further the payload sits from center, the harder the drone fights to stay level. Write that takeaway in your notebook before moving on.

9Build a data table: Position, Drift Distance (cm), Drift Direction.

Use a fresh battery for each trial. A weak battery drifts on its own. That contaminates your data.

I tested 4 positions

I used a fresh battery each time

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

Cheat Sheet - Why Offset Weight Causes Drift: TORQUE = force x distance from center. A 1-gram clip taped 5 cm forward of center creates a forward-tipping torque on the drone. The flight controller compensates by spinning the back motors faster than the front motors. This is called THRUST DIFFERENTIAL. CONTROL AUTHORITY = how much extra thrust the motors have available to counteract torque. Once payload offset exceeds control authority, the drone cannot correct fast enough and drifts. For your APEX 149: each motor can produce up to 30 g of thrust. With 4 motors, total ceiling is 120 g. Subtract the drone's own 31 g weight. That leaves 89 g of margin to redirect toward correcting offset payloads.

Why Does Offset Weight Drift?

1Open Gemini. Type: 'On a quadcopter, why does adding weight to the front cause the drone to drift forward instead of just sinking?'

2Read the AI's explanation. Look for the words 'torque' and 'thrust differential.' Write the AI's one-sentence answer in your notebook.

3Ask: 'What is the maximum offset payload my APEX 149 can handle before it cannot stabilize?' The AI will probably guess - record its number and compare to your data.

Engineers call this 'control authority' - how much corrective torque the motors have available. Once payload offset exceeds control authority, the drone is dead in the air.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Data

Position	Your Drift (cm)	Class Avg Drift (cm)
Front	fill	fill
Back	fill	fill
Left	fill	fill
Right	fill	fill
Center	fill	fill

1Which position had the smallest drift? Which had the largest? Why?

This is exactly the kind of data delivery-drone engineers gather to certify a payload bay design.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Aeronautical Engineer - Controlled Experiments

Aerospace Engineers Career Video

Check out this video to see a real professional in this role. Day-in-the-life of aerospace engineers across companies. Same controlled-experiment work you ran today.

Drone Test Engineer (Payload + Stability): What they do: design payload bays, certify drones to carry medical supplies, packages, cameras. Real LA employers: Skydio, Zipline, Amazon Prime Air. Entry pathway: BS Aerospace or Mechanical Engineering. Lots of internships at SoCal startups for sophomore-junior students. Salary band (Glassdoor LA 2024): entry 92,000 to 110,000. Mid-career 130,000 to 170,000. First step from where you sit today: build and document your APEX 149 work in a personal portfolio. Real employers want to see real builds.

Reflection

1Reflection 1: What surprised you about the data?

2Reflection 2: Where in your life is one variable secretly controlling everything else?

3Reflection 3: If you were certifying a delivery drone for medication transport, where would you place the payload?

I completed Day 7

I designed a controlled experiment

Built by Woven Learning & Technology✎ Edit

Day 8: Battery Drain and Flight Endurance

Aeronautical Engineer Pathway - Endurance Modeling

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

How Do Drones Really Fly?

Watch this short video to see why today's topic matters before we dive in. A thorough look at quadcopter aerodynamics. Useful background for your battery experiment.

Foundations - Battery Math

Lift, weight, and stability all depend on one thing: battery. A drone with a dead battery is a brick. Today you measure exactly how long YOUR drone can fly and what makes it shorter.

The Hook: Amazon Prime Air can only deliver to homes within 7.5 miles of their warehouse. Why? Battery. Each delivery drone has a max round-trip range based on battery capacity, payload weight, and weather. If they want to deliver further, they need a heavier battery, which means less payload, which is the whole problem. This is the engineering tradeoff at the heart of every electric flying machine on earth.

Foundations - Battery Math: Capacity = how much energy is stored (mAh = milliamp hours) Draw = how fast the motors use it (mA at hover) Flight Time (min) = Capacity / (Draw / 60) Your APEX 149 battery is small. Its real-world flight time is 9 to 10 minutes per fully-charged battery on a fresh charge - longer if you fly gently, shorter if you fly hard. Today you measure your APEX's actual hover time and compare it to the spec sheet.

1Predict: how many minutes can your drone hover before the low-battery warning?

I understand the battery flight-time equation

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: APEX 149 + controller + phone with Tspeed 7, FRESH fully-charged battery (verify at the charging station), spare battery (also fresh), stopwatch, paper clips for the payload trial, eyewear, taped flight zone, notebook.

Today's task with Pitsco Echo: measure how long your drone hovers across 4 battery charge levels (100%, 75%, 50%, 25%) so you can build a battery-life model and predict mission duration. Code Here - Pitsco Echo (block coding for APEX): Web app: https://echo.pitsco.com (Chrome on a laptop or tablet). Sign in with the class code your instructor shares. Open Tspeed 7 on your phone first, then connect Echo to the drone via the phone's Bluetooth. Drag blocks from the left palette into the canvas. Click RUN. The drone flies your code. The OFF switch on the drone or controller is your kill button.

The Hypothesis

1Predict: how long will your drone hover with NO payload? How long with a 2-gram payload?

The Build - Endurance Test

2Open the Drone Battery Predictor (above). Enter your drone's mass + 0g payload. The app gives you a predicted hover time at full charge. Record this as your NO-PAYLOAD prediction.

3Now enter the same drone mass + 2g payload. Note how predicted hover time drops. Record the percent difference - that's your PREDICTED penalty for adding 2g.

4Trial 1 (no payload): Build program: Takeoff (1 m) → Wait 600 s (10 min) → Land. Run. Start stopwatch at takeoff. Watch for auto-land or low-battery warning. Stop stopwatch when drone lands. Record total hover time.

5Swap to a fresh fully charged battery. Verify with the charging station LED.

6Trial 2 (with 2-gram payload taped to center): repeat the program. Time it. Record.

7Calculate the percent change in flight time when you added 2 grams. Was the drop bigger or smaller than you predicted?

Do NOT keep flying after the auto-land. Lithium batteries below 3.0 V can fail catastrophically. The drone protects itself; do not override.

I ran two endurance trials

I used fresh batteries

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Builds the Endurance Model

1Open Gemini. Type: 'For a APEX 149 with a 250 mAh battery and 1500 mA hover draw, build a table of expected hover time at payloads of 0, 1, 2, 3, and 5 grams. Show your assumptions.'

2Write the AI's predicted times next to your two actual data points. Does its model match the trend?

3Ask: 'What real-world factors would make actual flight time SHORTER than your prediction?' Read the answer.

Real factors: battery age, ambient temperature, propeller wear, motor calibration. The model is the ceiling. Reality is always under it.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Endurance Data

Payload (g)	Your Time (min:sec)	Class Avg (min:sec)	AI Prediction
0	fill	fill	fill
2	fill	fill	fill

1Calculate percent error between AI prediction and class average for each payload. Was the AI optimistic or pessimistic?

Real engineers always assume their drone will fly LESS than the spec sheet says. Plan for the worst case.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Aeronautical Engineer - Endurance Modeling

Energy Engineers Career Video

Check out this video to see a real professional in this role. Engineers who design battery + propulsion systems for everything that moves. The endurance math you did today.

Battery / Propulsion Engineer: What they do: design and certify the battery + motor combination on every electric vehicle, drone, and aircraft. Entry pathway: BS Electrical or Mechanical Engineering. Cal State LA EE program is a direct route. Internships at SpaceX, Joby, Archer, Lucid Motors, and Tesla all hire LA-area students. Salary band (Glassdoor LA 2024): entry 95,000 to 115,000. Mid-career 135,000 to 175,000. First step from where you sit today: take Cal State LA's free Saturday Engineering Workshops, sophomore year and up.

Reflection

1Reflection 1: How does the battery limit affect what kinds of drones can exist?

2Reflection 2: When have you 'run out of energy' on a project before finishing? What did you do?

3Reflection 3: If you had to design a delivery drone for medical supplies in LA, what is the FIRST tradeoff you would make: range or payload?

I completed Day 8

I have an endurance model

Built by Woven Learning & Technology✎ Edit

Day 9: Delivery Mission - Lift-to-Weight Synthesis

Aeronautical Engineer Pathway - Synthesis Day

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Three days ago you weighed the drone. Two days ago you tested payload position. Yesterday you measured battery endurance. Today you put all three together in a delivery mission.

How Wings Generate Lift (Recap)

Watch this short recap to bring back the key idea. Same Veritasium explainer you saw on Day 6 - this time keep an eye on how lift scales with weight, since today is the synthesis flight where lift-vs-weight is the whole game.

Foundations - Synthesis

Four days of data. Today you put it together. You will design YOUR drone's optimal flight profile - the payload it can carry the longest distance most stably.

The Hook: NASA's Ingenuity helicopter on Mars weighed 1.8 kg in Mars gravity (about 4 lbs back home). Mars atmosphere is 1 percent of Earth's, which means propellers have to spin 5 TIMES faster to generate any lift. Ingenuity barely had a lift-to-weight ratio above 1. NASA engineers called it 'flying through soup with cardboard wings.' It flew 72 times. Sometimes the math says NO. Engineers find a way.

Today combines: Lift-to-weight (Day 6) Payload position (Day 7) Battery endurance (Day 8) Your question: given a 2-gram package, what is the BEST flight profile (route, position, speed) to deliver it across your flight zone in the shortest time without crashing?

1Predict: which factor will dominate today - weight, position, or battery? Write your guess.

I understand today is synthesis of Days 5-8

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: APEX 149 + controller + phone with Tspeed 7, fresh battery, 2-gram payload (1 paperclip on a tiny tape pad), stopwatch, taped 3 m by 3 m flight zone with Start corner and Goal corner marked, eyewear, notebook. Whiteboard scoreboard at the front.

Today's task with Pitsco Echo: combine everything from this week (lift + position + battery) into ONE delivery mission - take off, carry payload across the room, drop it, return home. Code Here - Pitsco Echo (block coding for APEX): Web app: https://echo.pitsco.com (Chrome on a laptop or tablet). Sign in with the class code your instructor shares. Open Tspeed 7 on your phone first, then connect Echo to the drone via the phone's Bluetooth. Drag blocks from the left palette into the canvas. Click RUN. The drone flies your code. The OFF switch on the drone or controller is your kill button.

The Hypothesis

1Sketch your delivery mission in your notebook: start corner, goal corner, payload position on drone. Predict your delivery time.

The Build - Delivery Mission

2Open the Synthesis calculator (above). Enter your full mission setup: drone mass + 2g payload + a battery level you'll START at (say 80%). The app shows whether the drone can complete this mission. Status = HOVER ✅ means GO; status = BARELY HOVER means risky. Record the prediction.

3In Pitsco Echo, build: Takeoff → Fly Forward (3 m) → Land at Goal → Wait 2 s → Takeoff → Fly Backward (3 m) → Land at Start.

4Run with 2g payload taped to CENTER. Time the round trip. Record.

5Iterate: try the same mission with payload OFFSET to one side. Does it fly straighter or drift?

6Final attempt: optimize speed (slower turns, faster straights). Record best round-trip time.

7Compare your real round-trip time to the calculator's PREDICTION. Did the drone complete the mission as predicted? If status was BARELY HOVER and the drone made it - lucky. If status was HOVER ✅ but the drone crashed - what unmodeled factor caused the failure? (Wind? Pilot error? Battery sag mid-flight?)

Set up your team's DELIVERY SCOREBOARD now (different metric than Week 1's Showdown Scoreboard). Turn to a fresh page in your Woven notebook. Draw a 5-column table with these headers: Team | Total mass (g) | Payload position (center / offset) | Round-trip time (s) | Crash? (Y/N). Add a row for your team. The class also tracks all teams on the whiteboard - your notebook is your permanent record.

8Post your team's best round-trip time and mass on BOTH the class whiteboard AND your notebook DELIVERY SCOREBOARD. Then add the other teams' times to your notebook scoreboard as they post. End-of-class question: which team had the fastest delivery? What did they do differently?

Crashes count as DNF. A failed delivery is a failed mission - real Amazon drones have to land safely or it does not count.

I completed at least one full delivery mission

I optimized payload position

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Designs the Mission

1Open Gemini. Type: 'Design an optimal APEX 149 delivery mission: 2-gram payload, 3-meter round trip, fresh battery. Specify takeoff height, speed, and payload position. Show your reasoning.'

2Compare the AI's mission to yours. Did it pick the same speed? Same payload position? Same takeoff height?

3Run the AI's mission. Time it. Did it beat your best?

In real engineering, the AI is a starting point. Engineers tweak based on physical reality the AI does not know about - like the rough texture of the floor or the airflow from the AC vent.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Mission Times

Pair	Round-Trip Time (sec)	Crashes	Payload Position
Yours	fill	fill	fill
Class Best	fill	fill	fill

1What did the class winner do that you did not? Write one sentence.

This kind of head-to-head competitive testing is exactly how Amazon, Zipline, and Skydio refine their delivery drones. Iteration beats invention.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Aeronautical Engineer - Synthesis

Day at Work: Roboticist

Check out this video to see a real professional in this role. Dr. Marek Michalowski builds interactive robots. Mission design is exactly his daily work.

Mission Design Engineer: What they do: combine all the subsystems (battery, structure, control, payload) into a working aircraft mission. Half engineer, half project manager. Entry pathway: BS Aerospace or Mechanical Engineering with systems-engineering electives. Cal State LA + Cal Poly Pomona transfer. Salary band (BLS LA 2024): entry 90,000 to 110,000. Mid-career 130,000 to 175,000. First step: design and document a complete mission for a personal drone build, post it to GitHub or YouTube.

Reflection

1Reflection 1: Which day this week taught you the most? Why?

2Reflection 2: When have you combined separate skills into one final product?

3Reflection 3: Next week we make the drone fly itself. What concerns you about autonomous flight?

Week 2 Complete - pilots (Manual Flight): Five days of manual flight, payload tests, battery math, and lift-to-weight curves. You measured what flies and what doesn't with your own hands. Stand up. Read your single biggest data point out loud to the class. You finished Week 2. Group photo. Sign your notebook. See you next session.

I completed Week 2

I designed a delivery mission

Built by Woven Learning & Technology✎ Edit

Day 10: Echo Coding - Square and L-Shape Patterns

Autonomous Systems Engineer Pathway - First Block-Coded Flight

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

APEX 149 Coding Drone Tutorial

Carefully watch this tutorial to learn the steps before you try them. APEX's official walkthrough of the Pitsco Echo block coding environment. Watch this before you write your first program.

Foundations - Block Coding for Flight

Welcome back, pilot. Yesterday you flew the APEX with your hands on the sticks. Today the drone flies itself - because of the code you write.

The Hook: A Waymo self-driving car has 29 cameras, 6 radars, 5 lidars, and 4 GPS antennas. It generates 4 GIGABYTES of sensor data every minute it drives. All of that data feeds into ONE question: 'What should I do in the next 100 milliseconds?' Your APEX 149 is simpler. But the same idea: a program decides what the drone does, instead of you. Today you write that program.

MISSION BRIEFING - Apex Drone Navigator Challenge, Echo Edition: Welcome, drone pilots and coders. Your mission across Days 10 and 11: use block coding at https://echo.pitsco.com to program your APEX 149 to fly a pattern. Day 10: basic 'square-shaped' flight + L-shape pattern. Day 11: more complex designs using LOOPS and FUNCTIONS - a looped square, a zigzag, and a distance-increment mission where each leg gets longer. If your drone does not spin exactly 90 degrees on a turn block, adjust the time or speed parameter on the spin block. That tweaking IS the engineering.

Pitsco Echo is the block-coding tool for the APEX 149. You drag blocks like 'Take Off', 'Forward', 'Left 90 degrees', 'Land' into a stack. The drone runs them in order. Key blocks you will use today: - TAKE OFF (lifts to default hover height) - FORWARD (distance in inches) - TURN LEFT / TURN RIGHT (degrees) - LAND The whole program is the recipe for the flight.

1Talk to your partner: if you wanted the APEX to fly a perfect square, what blocks (in what order) would you need?

2Predict: if your drone needs to fly a 60 cm square at full throttle, will you need the same number of blocks as a 30 cm square?

I can name 4 Echo block types

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Safety: eyewear ON. Stay in the taped flight zone. Crash protocol still applies even with autonomous code - keep the controller close as a manual override. The OFF button on the drone or controller stops everything.

Materials per pair: APEX 149 + controller + laptop or tablet with https://echo.pitsco.com open in Chrome, fresh battery, painters tape (mark a Start spot on the floor), ruler, eyewear, taped 3 m by 3 m flight zone, notebook. Charging station at instructor table.

Today's task with Pitsco Echo: switch from manual flight to AUTONOMOUS code. You write the block code, hit run, and the drone flies itself. Today's first autonomous mission: a square pattern. Code Here - Pitsco Echo (block coding for APEX): Web app: https://echo.pitsco.com (Chrome on a laptop or tablet). Sign in with the class code your instructor shares. Open Tspeed 7 on your phone first, then connect Echo to the drone via the phone's Bluetooth. Drag blocks from the left palette into the canvas. Click RUN. The drone flies your code. The OFF switch on the drone or controller is your kill button.

APEX 149 Coding Drone Tutorial

Carefully watch this tutorial to learn the steps before you try them. APEX's official walkthrough of the Pitsco Echo block coding environment. Watch this before you write your first program.

The Hypothesis

1Sketch the square pattern in your notebook. Mark the corners A, B, C, D. Predict: how many blocks total will your program need (Take Off + 4 Forwards + 4 Turns + Land = 10 blocks)?

The Build - Square Pattern

2Open https://echo.pitsco.com on your laptop. Sign in with the class code your instructor shares.

3Drag the blocks for a square in this order: Take Off then Forward 24 inches then Turn Right 90 then Forward 24 inches then Turn Right 90 then Forward 24 inches then Turn Right 90 then Forward 24 inches then Land.

4Place the drone on the marked Start spot. Click RUN. Watch the flight.

5Did it close the square? Measure the gap from where it landed to the Start spot. Record.

6Adjust if needed: if the drone undershoots, increase Forward distance. If it does not turn 90 degrees cleanly, adjust the turn block (Tip: Echo lets you tweak the turn-block timing or speed if 90 is not exactly 90).

7Re-run. Measure new gap. Goal: within 15 cm of Start in 3 attempts.

The Build - L-Shape Pattern

8Now write a program for an L-Shape: Take Off then Forward 36 inches then Turn Left 90 then Forward 24 inches then Land.

9Run it. Did it land at the corner of the L?

Doesn't have to be perfect on the first run. Real engineers iterate. Tweak the numbers, re-run, observe.

My APEX flew a square within 15 cm of Start

My APEX flew an L-shape

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Critiques Your Block Code

1Take a screenshot of your block code in Echo. Paste the screenshot (or describe the block sequence in plain English) into Gemini. Type: 'Here is my block-coded drone flight. What is one thing I could simplify or improve?'

2Read the AI's suggestion. Common AI answers: 'use a Repeat loop instead of writing 4 turns separately' (preview of tomorrow), or 'set the turn duration manually for more accurate 90 degrees'.

3Apply the AI's suggestion (if it makes sense to you). Re-run. Did it improve?

AI cannot run your drone, but it can READ your block sequence and spot inefficiency. Real software engineers do this every day with AI code review.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Square Accuracy

Square-Pattern Gap to Start (cm)

Pair	Attempt 1	Attempt 2	Attempt 3	Tweak made
A
B
C
D

1Which pair had the smallest final gap? What did they tweak?

Closing the loop within 15 cm is REAL precision for a small drone. Pro drone shows are choreographed to within 5 cm. Your APEX is not far behind.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Drone Software Engineer

Day in the Life of an Embedded Engineer

Check out this video to see a real professional in this role. Embedded engineers write the low-level autopilot code drones run. Same blocks you wrote today, in C and Python.

Drone Software Engineer: What they do: write the autopilot code that flies commercial drones, military UAVs, and consumer hobby drones. The code that you ran today is the same kind of code these engineers write. Entry pathway: BS Computer Science or Computer Engineering. Cal State LA's CS BS plus drone-club projects. Internship pipeline: Skydio, Anduril, AeroVironment - all SoCal companies hiring CS students. Salary band (Glassdoor LA 2024): entry 100,000 to 130,000. Mid-career 145,000 to 200,000. First step from where you sit today: keep your Echo programs in a portfolio. Apply to summer programs at JPL, Lockheed, or local hackathons.

Spotlight: Dr. Ayanna Howard: Dr. Howard (1966 to present) is a roboticist who worked at NASA's Jet Propulsion Laboratory. She developed SmartNav - the software that lets Mars rovers move autonomously. The Curiosity rover does not need every move radioed from Earth because of the code engineers like her wrote. She is now a pioneer in Human-Robot Interaction - teaching robots to help and learn from people. The block code you wrote today is the simplest version of what she designed for Mars.

Reflection

1Reflection 1: what surprised you about how the AI saw your code?

2Reflection 2: what is the difference between code that reads correct and code that flies correct?

3Reflection 3: tomorrow we add LOOPS. What do you predict a loop will let you do?

My APEX flew a square autonomously

I logged the gap-to-Start data

Built by Woven Learning & Technology✎ Edit

Day 11: Echo Coding - Loops, Zigzag, and Distance Math

Autonomous Systems Engineer Pathway - Repetition and Variables

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

APEX 149 Coding Drone Tutorial (Refresher from Day 10)

Re-watch this short refresher to remember what you learned earlier. Same APEX coding tutorial. Re-watch the loop section if you need a refresher.

Foundations - Loops

The Hook: In 2010 a single Tesla manufacturing line had 90 robotic arms. Each arm did the SAME 30 motions on every car, 24 hours a day. The arms did not have unique programs - they had ONE program, written inside a LOOP. Loops are how programs scale from doing one thing to doing 10,000 things. Today you teach your drone its first loop.

Yesterday you wrote 4 Forward + 4 Turn blocks for a square. That is 8 blocks of repetition. Today you write ONE Forward + ONE Turn inside a REPEAT 4 TIMES block. That is 3 blocks total. Same square. Same drone path. Different code. Less repetition. The skill is recognizing patterns and replacing them with loops.

1Talk to your partner: name 3 things in your daily life that are loops.

2Predict: a square pattern with a LOOP block versus a square pattern WITHOUT a loop - will they fly the same? Or will the loop change the flight path?

I can write a square as a loop

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: APEX 149 + controller + laptop with https://echo.pitsco.com open, fresh battery, painters tape (mark Start), eyewear, flight zone, notebook.

Today's task with Pitsco Echo: replace 4 separate move blocks with a LOOP - so the drone can do more with less code. You will also vary distance with a hopscotch pattern. Code Here - Pitsco Echo (block coding for APEX): Web app: https://echo.pitsco.com (Chrome on a laptop or tablet). Sign in with the class code your instructor shares. Open Tspeed 7 on your phone first, then connect Echo to the drone via the phone's Bluetooth. Drag blocks from the left palette into the canvas. Click RUN. The drone flies your code. The OFF switch on the drone or controller is your kill button.

APEX 149 Coding Drone Tutorial (Refresher from Day 10)

Re-watch this short refresher to remember what you learned earlier. Same APEX coding tutorial. Re-watch the loop section if you need a refresher.

The Hypothesis

1Predict: a square coded with a loop and a square coded without a loop - will they fly differently? Or fly the same?

Build 1 - Square with a Loop

2Open Echo. Build: Take Off then REPEAT 4 TIMES { Forward 24 inches then Turn Right 90 } then Land.

3Run. Measure gap to Start. Compare to yesterday's square gap. Same? Better? Worse?

Build 2 - Zigzag Pattern

4Build: Take Off then REPEAT 3 TIMES { Forward 18 inches then Turn Right 45 then Forward 18 inches then Turn Left 45 } then Land.

5Run. Watch the zigzag path. Adjust the angles or distances if it does not look like a clean zigzag.

Build 3 - Distance Increment Challenge

6Mission: take off, fly forward, return. Repeat 3 times. Each time, increase the forward distance by 12 inches: 12, then 24, then 36.

7Sketch the flight path in your notebook before coding. Then build it in Echo (you may need to write the 3 forward-and-back sequences out separately, or use a more advanced loop with a variable - try both).

8Run. Watch the drone make 3 progressively longer out-and-back trips. Check that it returns to the Start spot each time.

Real engineers measure floor space first. 12 + 24 + 36 = 72 inches forward at the longest leg. Make sure your zone is bigger than that. Adjust the increments down if needed.

I flew a looped square

I flew a zigzag pattern

I flew the distance-increment mission

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Refactors Your Code

1Take your distance-increment program (3 separate forward-and-back sequences). Paste it (or describe it) into Gemini. Ask: 'Can this program be rewritten using a loop with a variable that increments by 12 each time? Show me the cleaner version.'

2Read the AI's refactored code. Translate it into Echo blocks if you can.

REFACTORING is when you keep the program's behavior the same but make the CODE shorter or clearer. Real engineers refactor weekly. AI is great at this kind of work.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Loop Efficiency

Block Count - With vs Without Loops

Pair	Square (no loop)	Square (with loop)	Zigzag	Distance Increment
A
B
C
D

1What is the average block-count savings when using a loop?

Loops are how a single coder writes a program that does 1,000 things. Real autopilots use thousands of nested loops every second.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Embedded Systems Engineer

Day in the Life of an Embedded Software Engineer

Watch this video to see what a real day in this career actually looks like. A working embedded engineer's day. Loops and variables are her bread and butter.

Embedded Systems Engineer: What they do: write the low-level code that runs inside cars, drones, medical devices, and IoT sensors. Loops are their bread and butter. Entry pathway: BS Computer Engineering or Electrical Engineering. Cal State LA's CompE BS is direct. Internships at Northrop, Boeing, Honeywell, all SoCal-based. Salary band (Glassdoor LA 2024): entry 100,000 to 125,000. Mid-career 140,000 to 190,000.

Reflection

1Reflection 1: where in your life have you noticed a pattern that helped you do something faster?

2Reflection 2: when does a loop make code clearer? When does it make code harder to read?

3Reflection 3: tomorrow we run a real mission - search and rescue on Catalina Island. What kind of code might you need?

I built and flew a looped program

I have notebook-logged the block-count savings

Built by Woven Learning & Technology✎ Edit

Day 12: Catalina Mission Phase 1 - First Sighting

Autonomous Systems Engineer Pathway - FPV Flight + Visual Acquisition

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Yesterday you wrote your first autonomous flight - a square pattern. Today you write loops and zigzags so the drone can do more with less code.

Santa Catalina Island - Aerial 4K View

Watch this short video to see why today's topic matters before we dive in. Real aerial drone footage of the actual island where today's mission takes place. This is your operational area.

Binary Numbers - Refresher

Re-watch this short refresher to remember what you learned earlier. Techquickie's 4-minute explainer. Watch this if you do not remember how binary works - you will need it to decode the clue.

Foundations - First-Person View (FPV)

- we have a real mission today.

The Hook: A research drone has crashed on an unstable crater wall on Catalina Island. Before its main power failed, its e-ink screen froze, displaying a binary code message - our first vital clue.

MISSION BRIEFING - Catalina Code Recovery, Phase 1: First Sighting: A research drone has crashed on an unstable crater wall on Catalina Island. Before its main power failed, its e-ink screen froze, displaying a binary code message - our first vital clue. The section of the crater wall with the first code is in a known zone. You can see the general area from your launch point. Your Objective: - From your designated launch area, pilot your drone toward the crater wall where the first binary code is displayed. - Even though you might see the target with your own eyes, your primary goal is to practice steady flight and clear visual acquisition through your drone's camera. - Use the live video feed in the Tspeed 7 app to get a stable, readable view of the binary code. - Write down the code or capture a clear photo through the app. - Practice smooth maneuvers and controlled hovering to ensure the code is legible in your video feed. - Safely return your drone to the launch point. Challenge Advisory: that crater wall is fragile. Unsteady flight or hovering too close kicks up dust and makes the code difficult to read on your screen. Focus on precise control.

FPV = First-Person View. You see what the drone sees, through its camera, on your phone screen. Your APEX has an onboard camera that streams live video to the Tspeed 7 app. Today you fly with EYES on the drone (visual flight) but use the camera FEED to capture data the drone can see better than you can. This is how real search-and-rescue drones work: a pilot flies the drone close to a survivor, and the camera reads what the human eye cannot.

1Predict: how steady will your hover need to be for the binary code on the e-ink screen to be readable on your phone?

I understand the FPV concept

I can decode binary

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: APEX 149 + controller + phone with Tspeed 7, fresh battery, 1 printed binary clue at hover height (instructor pre-tapes), painters tape Start spot, eyewear, flight zone, notebook with binary decoding chart.

Today's task with Tspeed 7 (FPV Mode): fly with the drone's onboard camera as your only view. The screen IS your cockpit - first day of seeing through the drone's eyes. Code Here - Tspeed 7 FPV App: Today is the FIRST DAY you fly the drone via the PHONE APP, not the controller. Per the manual: 'The controller will not work if the drone is connected to your device. The app control will not work if the drone is paired to a controller.' Pick one. Today: app. Steps: 1. Power off the controller (so the drone is not paired to it). 2. Power on the drone. The blue LED will flash. 3. On your phone: open Settings > WiFi. Connect to the network 'Tspeed7-XXXXX' (the X's are unique to your drone, printed on the drone underside). Your phone will say 'No Internet' - that is correct; you are connected to the drone, not the internet. 4. Open the Tspeed 7 app. Tap CONTROL. The live camera view appears with virtual joysticks on screen. iOS: https://apps.apple.com/in/app/tspeed-7/id1575505851 Android: https://play.google.com/store/apps/details?id=com.apex.at149 On the screen: virtual LEFT joystick (altitude + yaw), virtual RIGHT joystick (forward/backward + roll). Same map as the physical controller. SCREENSHOT button bottom right - use it to capture the binary clue.

How to Fly FPV Drones - The 4 Stick Controls

Carefully watch this short video to learn FPV flying BEFORE you take off. The 4 stick controls (throttle, yaw, pitch, roll) on Tspeed 7's virtual joysticks work the same as on any FPV drone. Once you have the muscle memory, the app is just the screen showing what the drone sees.

Safety: eyewear ON. Stay in your taped flight zone. The 'crater walls' (printed binary code on cardstock) are NOT to be flown into - hover 30 cm away minimum. Crash protocol: LEFT STICK ALL THE WAY DOWN.

Mission Phase 1 - Visual Sighting

1Pre-flight: connect drone to controller. Connect phone to drone WiFi. Verify the live camera feed is working before takeoff.

2Set the phone where pilot OR co-pilot can see the screen. Decide who watches the screen and who watches the drone. Communicate.

3Take off from the marked Start. Fly slowly toward the binary clue on the wall. Hover 30 cm away.

How to capture and recover the photo (Tspeed 7 specifics): 1. The Tspeed 7 app has a SCREENSHOT button in the bottom-right corner of the live FPV view. Tap it once when the binary code is sharp and centered in the frame. The button briefly flashes white to confirm. 2. The screenshot saves to your phone's Photos app (iOS) or Gallery / Files (Android). It is a regular .jpg image you can email, AirDrop, share to Drive, or upload directly. 3. For Phase 3 today: open https://gemini.google.com (sign in with your school Google account). Tap the paper-clip / image-upload icon in the prompt box. Pick the screenshot you just captured. Then type your decode prompt. If the screenshot is blurry or off-center, fly back, re-hover steady, and capture again. Better to retry than to upload a useless image.

4Co-pilot: watch the live feed on the phone. When the binary code is sharp and readable, take a screenshot. Confirm out loud: 'Got it.'

5If the screenshot is blurry or the code is unreadable, reposition the drone (steady hover, slightly different angle, more light) and re-screenshot.

6Pilot: when the co-pilot confirms, fly the drone back to Start. Land.

Quick binary decoder reference: each 8-bit binary number = 1 ASCII character. Examples: A=01000001, B=01000010, C=01000011, S=01010011, O=01001111. To decode, split your binary into 8-digit groups (separated by spaces or every 8 digits), then look up each group on a reference chart - or paste the whole binary string into this online converter: https://cryptii.com/pipes/binary-to-text/ (paste left side, read text on right side). Same site works for tomorrow's Day 13 mission too. Verify your hand-decode against the converter - if they disagree, recheck your spacing.

7Together: decode the binary code into letters or numbers. Write the decoded message in your notebook.

Doesn't have to be perfect on the first run. If the screenshot is blurry, take off again. Real search-and-rescue pilots routinely re-attempt - that is the work.

I captured a clear screenshot of the binary clue

I decoded the binary into a message

My drone is back at Start, intact

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Reads the Image

How To Upload an Image to Google Gemini AI

Watch this short tutorial first if you have never uploaded an image to Gemini. Same flow on phone or laptop: tap the paper-clip / image icon in the prompt box, pick your binary screenshot, then type your decode prompt below the image.

1Open Gemini. Upload your binary screenshot. Type: 'What binary code is on this e-ink screen? Decode it to ASCII.'

2Compare the AI's decoded message to your hand-decoded message. Did they agree?

3If they disagreed, who is right? Re-check your manual decoding step by step.

This is exactly how modern search-and-rescue teams work. The drone captures the image. The pilot decides whether the AI's reading is trustworthy. The pilot has the final call.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Mission Log

Phase 1 Mission Outcomes

Pair	Decoded message	Attempts	AI agreed?
A
B
C
D

1Did everyone decode the same message? If not, why?

Real search-and-rescue requires multiple confirmations. One drone, one pilot, one screenshot = not enough. Multiple teams cross-checking = trustworthy.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Search and Rescue Drone Pilot

9 Growing Jobs for Drone Pilots in 2025

Check out this video to see a real professional in this role. Search-and-rescue is one of the fastest-growing branches of commercial drone work.

Search and Rescue Drone Pilot: What they do: fly drones into wildfires, mountainsides, collapsed buildings, and flooded areas to locate survivors. They work for fire departments, the Coast Guard, and state emergency management. Entry pathway: FAA Part 107 license (you can take the test at 16). California also has search-and-rescue volunteer pilot programs. CalFire actively recruits drone pilots. Salary band: volunteer-track until 18, then 50,000 to 90,000 for full-time public-safety drone pilots. Specialty SAR pilots can hit 100,000+.

Reflection

1Reflection 1: who took the lead in your pair - pilot or co-pilot? Why did that work?

2Reflection 2: when has a teammate seen something you missed?

3Reflection 3: tomorrow you fly BLIND - no line of sight, only the camera feed. What concerns you?

I completed Phase 1

I have the decoded message in my notebook

I know my role for tomorrow

Built by Woven Learning & Technology✎ Edit

Day 13: Catalina Mission Phase 2 - Blind FPV Retrieval

Autonomous Systems Engineer Pathway - Camera-Only Navigation

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Today's goal: fly the drone using ONLY its onboard camera (no looking at the drone). Find a hidden binary code on the wall and decode it.

How Self-Driving Cars See

Watch the video to set up today's thinking - this is the real-world story behind the lab. How autonomous systems navigate using camera feeds when the human cannot see directly. Same skill, smaller drone.

Foundations - Trusting the Feed

The Hook: Intel indicates a SECOND, even more critical binary code on a different, remote section of the Catalina crater wall. This time, the target area is NOT directly visible from your piloting station - you fly entirely on the drone's camera feed.

MISSION BRIEFING - Catalina Code Recovery, Phase 2: Blind Data Retrieval: Excellent work yesterday, pilots. Intel now indicates a SECOND, even more critical binary code on a different, remote section of the Catalina crater wall. This time, the target area is NOT directly visible from your piloting station. Your Objective: - Launch your drone from a position where you cannot see the target wall with your own eyes (the instructor will set up a screen or designate a hidden location). - This is a First-Person View (FPV) ONLY challenge. You rely entirely on your drone's live video feed through the Tspeed 7 app. - Using only the drone's camera, navigate to find the new binary code on the designated crater wall section. - Once located, capture a clear screenshot of this second code. - Safely navigate your drone back to its launch point using FPV only. Critical Challenge: flying blind using only your drone's camera demands patience, smooth control, and clear communication with your team. Trust your drone's perspective.

When you cannot see the drone, the camera feed IS your reality. The screen shows you a narrow view (the drone's camera has a limited field of view). Things to your left and right are invisible. Three FPV survival rules: 1. SLOW. Move slowly. Misreading the screen at speed = crash. 2. ROTATE TO LOOK. If you can't find the target, rotate the drone (yaw) to scan the room. 3. ALWAYS KNOW WHERE HOME IS. Mentally track which direction is back to the launch point.

1Predict: when you fly blind, will you crash more, the same, or less than when you fly with line of sight?

I understand FPV-only flight

I know the 3 survival rules

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: APEX 149 + controller + phone with Tspeed 7, fresh battery, 1 printed binary clue at hover height in the hidden zone, painters tape Start spot, screen or divider blocking line of sight, eyewear, flight zone, notebook.

Today's task with Tspeed 7: extend yesterday's FPV flying into a TRUE blind mission - line of sight to the drone is blocked by a screen, so the camera feed is your ONLY information. Code Here - Tspeed 7 App (FPV Mode): Power OFF the controller (so drone is not paired to it). Power on the drone. On your phone: Settings > WiFi. Connect to the network 'Tspeed7-XXXXX' (printed on the underside of YOUR drone). Phone says 'No Internet' - that's correct. Open Tspeed 7 app. Tap CONTROL. Live camera feed appears with virtual joysticks on screen. iOS: https://apps.apple.com/in/app/tspeed-7/id1575505851 Android: https://play.google.com/store/apps/details?id=com.apex.at149

Safety: eyewear ON. Spotter standing OUTSIDE the FPV-only zone watches for runaway drones and calls 'STOP' if needed. Pilot is forbidden from peeking. Crash protocol: LEFT STICK ALL THE WAY DOWN.

Mission Phase 2 - Blind Retrieval

Why this app: today you fly BLIND - eyes on the screen only - so the drone's camera has to see a small printed binary code clearly. If the room is too dark, the camera reads garbage. The ArUco Lux Simulator (below) lets you slide the lux value low and watch detection accuracy collapse - same physics as your live mission. Run it for 2 minutes BEFORE you take off so you know what 'enough light' looks like.

1Open the ArUco Lux Simulator (above). Slide the lux value DOWN from 1000 → 500 → 200 → 100 → 50. At each step, watch detection accuracy. Find the lux threshold where detection drops below 80% - that's your minimum lighting requirement for today's blind FPV mission.

Lux Pre-Check (5 min)

Lux check on your room: use a phone Lux meter app to measure the actual room. If it reads above 200 lux, you are clear to fly. Below 200 - turn on more lights or move closer to the window.

2Use a phone Lux meter app (or estimate). Is the room above 200 lux? If not, ask the instructor to turn on more lights.

3Decide roles: PILOT (sticks, sees only screen) and COMMS (watches the spotter for safety calls, also tracks 'mental map' of where home is).

4PILOT: stand behind the screen so the drone is not visible. Phone running Tspeed 7 in hand (drone is paired to phone, NOT controller, today). COMMS: same side as PILOT, watches the spotter for safety calls and tracks 'mental map' of where home is.

5Take off slowly from Start. Watch ONLY the screen.

6Search for the binary clue: yaw slowly, scan the room. When the clue appears in the camera view, hover and steady.

Photo workflow (refresher from Day 12): tap the SCREENSHOT button in the bottom-right of Tspeed 7's FPV view when the binary clue is sharp. The image saves to your phone's Photos / Gallery. Today you will decode the binary by hand together with yesterday's message - no Gemini upload needed unless you want to verify.

7Capture screenshot of the second binary clue. COMMS confirms 'got it'.

8Navigate home: yaw the drone to face the launch direction (use the mental map). Fly forward to home. Land.

Binary decoder refresher: same converter as Day 12 - https://cryptii.com/pipes/binary-to-text/ - paste the binary on the left, read the ASCII on the right. Combine today's decoded word with yesterday's to see the full mission message.

9Decode the binary. Together with yesterday's message, what does the combined clue say?

If your drone disappears from the camera view (flew under a desk, etc.) - call SPOTTER. Do not panic-stick. The spotter walks to the drone and recovers it physically.

I flew the drone using ONLY the camera feed

I captured the second binary clue

I decoded the combined message

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Suggests Tracking Improvements

1Open Gemini. Type: 'My drone follows visual targets but loses them when the camera pans. What 3 algorithm improvements would let it lock on to a target and follow it automatically?'

2Read the AI's suggestions. Common answers: predict motion (Kalman filter), increase frame rate, widen search box, color thresholding.

3Discuss as a class: which of these would matter most for our binary-clue mission?

This is the work autonomy engineers do every day. They generate ideas, evaluate them, and ship the ones that work. You just did real-world autonomy engineering.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Week 2-3 Synthesis

Mission Phase Outcomes

Pair	Phase 1 message	Phase 2 message	Combined meaning
A
B
C
D

1What is the full combined message? Did all pairs converge?

2Share your team's final decoded message with the class. The instructor will collect each team's message on the front board so we can compare across pairs.

You just completed 9 days of aerodynamics + autonomy. You programmed manual flight, autonomous patterns, FPV missions, and decoded real binary clues. You know more about drones than 99 percent of US high schoolers.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Autonomous Systems Engineer

Day in the Life - Boston Dynamics Robotics Engineer

Watch this video to picture yourself in this career 5 to 10 years from now. Closing the loop between camera and motors is exactly what Boston Dynamics engineers do every day.

Autonomous Systems Engineer: What they do: integrate perception (camera), planning (algorithm), and control (motors) into one working autonomous system. The 'glue' engineer between AI teams and hardware teams. Entry pathway: BS Computer Engineering or Robotics. Cal State LA + transfer to USC's MS in Robotics or UCLA's CS Vision concentration. Salary band (Glassdoor LA 2024): entry 115,000 to 140,000. Mid-career 165,000 to 220,000.

Spotlight: Dr. Gladys West: Dr. West (1930 to 2026) was a mathematician at the US Naval Surface Warfare Center who did the foundational math for GPS - the system every autonomous drone, car, and plane uses to know where it is. When you flew your APEX with a heading and distance today, you were using technology that Dr. West invented. For most of her career, her name was barely known. She was inducted into the Air Force Hall of Fame at age 88. Brilliance is sometimes invisible until someone bothers to look.

Reflection

1Reflection 1: what was hardest about flying without seeing the drone?

2Reflection 2: when in your life have you had to trust something you could not see?

3Reflection 3: next week we trade drones for stethoscopes. What does autonomy look like in medicine?

Week 3 Complete - pilots (Autonomy + Catalina): Four days of autonomous flight, programming the drone to fly itself, and recovering binary clues from a search-and-rescue mission. Phase 1 visual sighting + Phase 2 blind FPV retrieval - both decoded. Stand up. Group photo with your decoded message held up. You finished Week 3. Group photo. Sign your notebook. See you next session.

I completed Phase 2 of the Catalina mission

I shared my team's decoded message with the class

I completed Week 2-3 of the program

Built by Woven Learning & Technology✎ Edit

Day 14: Vital Signs - Reading the Body in Real Time

EMT Pathway - Reading the Body in Real Time

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Today's goal: measure your own heart rate, breathing rate, and oxygen level. Then compare what those numbers mean for a real EMT making a call.

How the Heart Pumps Blood

Watch this short video to see why today's topic matters before we dive in. TED-Ed's clean animation of the heart's pumping cycle. The vital signs you measure today come from this engine.

Foundations - The Four Vitals

Robots make decisions based on sensor data. So do paramedics, ER nurses, and ICU doctors. This week you become the sensor. You measure heart rate, breathing rate, blood pressure, and oxygen - the four vital signs that define life.

The Hook: An EMT in LA County responds to a call: 'patient, age 17, unconscious.' She is on scene in 7 minutes. The first 30 seconds are not about diagnosis. They are about VITALS. Heart rate. Breathing rate. Pulse oximetry. Blood pressure. Those four numbers tell her if the patient is dying or stable. Four numbers. Every code. Every shift. For her entire career.

HR (Heart Rate): normal 60-100 beats per minute. Below 60 = bradycardia. Above 100 = tachycardia. RR (Respiratory Rate): normal 12-16 breaths per minute. Below 12 = bradypnea. Above 20 = tachypnea. SpO2 (Oxygen Saturation): normal 95-100 percent. Below 90 percent = hypoxia. BP (Blood Pressure): normal under 120/80. Above 140/90 = hypertension. These are the baselines every healthcare worker memorizes.

1Predict: what do you think YOUR resting heart rate is right now?

I can name the four vital signs

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 fingertip pulse oximeter (e.g., Zacurate brand), stopwatch, notebook, an open ~2 m x 2 m area for jumping jacks (cleared bench / hallway). Optional: BP cuff (manual or automatic) - one shared per 4 pairs. All cases anonymized; if a student has a known cardiac condition they can sit out the activity.

Pulse Oximeter — Nursing Skill

Check out this video to learn the steps before you do them yourself. Exactly how to place a pulse oximeter. Use this technique on yourself and your partner today.

The Hypothesis

1Predict your resting HR, RR, and SpO2. Write the 3 numbers.

The Build - Measure Yourself

2Take your real measurements: pulse oximeter on finger (HR + SpO2), and count breaths for 30 seconds × 2 (RR). Open the Vital Signs Interpreter (above). Enter all 3 numbers. The app shows whether each is in NORMAL, BORDERLINE, or ALARM range - and what an EMT would say next.

3Compare to your prediction. Were any of your guesses off by more than 10%? Write what you learned about your own resting baseline - this matters for the rest of the medical week.

4Sit quietly for 1 minute. Place the pulse oximeter on your index finger. Wait 10 seconds for stable reading. Record HR and SpO2.

5Without thinking about it, count your breaths for 30 seconds. Have your partner watch a stopwatch. Multiply by 2 = RR. Record.

6Compare to your predictions. Were you in the normal range?

7Now do 1 minute of jumping jacks (or stand-and-reach if you cannot do JJs). Sit immediately. Re-measure HR and SpO2. Record.

8Wait 2 minutes. Re-measure. Record this 'recovery' value.

Station 4 - Manual Pulse Palpation (Radial + Carotid)

Why this matters: oximeters fail. Batteries die, sensors get dirty, fingers are too cold. Every EMT knows how to take a pulse with two fingers - the same skill humans have used for 3000 years. Today you practice on yourself, then check your manual count against the oximeter.

9Find your RADIAL pulse: turn your left palm up. With the index + middle finger of your RIGHT hand, press lightly on the thumb-side of your left wrist, ~2 cm above the wrist crease. You should feel a steady thump. (If not, slide your fingers around 1 cm - it varies by person.)

10Count beats for 30 seconds, multiply by 2 = beats per minute. Record. Compare to your oximeter reading from Station 1. They should match within 5 bpm.

11Now find your CAROTID pulse: tilt your head slightly back. With light fingertip pressure (NOT pushing hard), press the side of your neck about 2 cm below your jaw, on either side of your trachea. Steady thump. Count for 30 seconds, double, record.

Station 5 - Active Recovery Curve

Why this matters: cardiovascular fitness is measured by how QUICKLY your heart returns to resting after exercise. Athletes recover in 60 seconds. Sedentary adults take 3+ minutes. This is the same metric Olympic doctors track - and you can measure it on yourself in 4 minutes.

12Take your RESTING HR with the oximeter. Record.

13Do 30 JUMPING JACKS at full speed. Stop. Sit. IMMEDIATELY measure HR with the oximeter. Record.

14Continue measuring HR every 30 seconds for 3 minutes (so 6 readings: 0, 30, 60, 90, 120, 150 sec post-exercise). Plot on a quick graph in your notebook.

15Look at your curve. How long until you got within 10 bpm of resting? Faster recovery = better cardiovascular fitness. Anything under 90 seconds is athlete-territory; 90-180 sec is healthy adult; 180+ sec means more cardio in your week would help.

If your SpO2 reads below 90, raise your hand. The oximeter could be miscalibrated, OR you could need attention. Either way the instructor checks it.

I measured my resting and active vitals

I tracked recovery

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI as Pattern Recognizer

AI does NOT diagnose. AI recognizes patterns. A diagnosis requires a licensed physician with a full patient history. We use AI today as a teaching tool only.

1Open Gemini or Claude. Type: 'A patient has HR 130, RR 28, SpO2 88. List 5 possible patterns these vitals are consistent with. Do not diagnose - just describe what conditions show this pattern.'

2Read the answer. Write the 5 patterns in your notebook.

3Now ask: 'What single additional measurement would help narrow these down?' Read the answer.

Real ER triage works exactly like this. Vitals do not give a diagnosis. They give a PRIORITY (red, yellow, green) and a list of likely paths to investigate.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Vital Signs

Vital	Class Min	Class Avg	Class Max	Normal Range
HR (rest)	fill	fill	fill	60-100
RR (rest)	fill	fill	fill	12-16
SpO2 (rest)	fill	fill	fill	95-100

1Was the class average inside the normal range? Were there outliers? Why might that be?

Athletes often have HR below 60 (resting trained heart). High caffeine pushes HR up. The 'normal range' is a starting point, not a verdict.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: EMT - Reading the Body in Real Time

Emergency Medical Technicians and Paramedics Career Video

Watch this video to picture yourself in this career 5 to 10 years from now. EMTs make the four-vital-signs read every shift, every call, every patient.

EMT (Emergency Medical Technician): What they do: first responder on every 911 medical call. Stabilize, assess, transport. Entry pathway: 14-20 weeks of EMT class + state cert. Mt SAC College, LA City College, and East LA College all offer the program. Often FREE for low-income students. Salary band (BLS LA 2024): entry 50,000 to 65,000. Mid-career 70,000 to 90,000. First step from where you sit today: shadow an EMT through the LA County Fire Explorer Program (high school sophomore eligible).

Reflection

1Reflection 1: What surprised you about your own vitals?

2Reflection 2: When have you had to make a fast judgment with limited information?

3Reflection 3: How does data change a real-time decision?

Spotlight: Dr. Patricia Bath: Dr. Bath (1942 to 2019) was a Black female ophthalmologist who invented the Laserphaco Probe in 1986 - a laser device that revolutionized cataract surgery and restored sight to millions worldwide. She was the first Black woman to receive a medical patent. She also founded the American Institute for the Prevention of Blindness with the principle that 'eyesight is a basic human right.' Her work bridges medicine, optical engineering, and access. The vital-signs work you do today is the bedrock of the larger bridge she built.

I measured my own vitals

I understand the normal ranges

Built by Woven Learning & Technology✎ Edit

Day 15: Blood Pressure - The Pump Pressure Story

Cardiology Tech Pathway - Korotkoff Sounds and Cuff Technique

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Today's goal: learn the manual blood pressure technique - cuff, stethoscope, gauge - and take 4 readings on yourself across rest, activity, and recovery.

How the Heart Pumps Blood (Recap)

Watch this short recap to bring back the key idea. Same heart, this time we focus on how each beat creates the pressure you will measure.

Foundations - Two Numbers

Yesterday: heart rate, breathing, oxygen. Today: the most underrated vital sign on the body - BLOOD PRESSURE. It is the pressure of the pump that keeps you alive.

The Hook: A 50-year-old man walks into a clinic with no symptoms. Routine check. BP reads 180/110. He does not feel anything. He has had this number for 6 months. If he leaves untreated, he has a 50 percent chance of a stroke or heart attack in the next 5 years. BP is called the silent killer because by the time you feel it, the damage is done.

Systolic (top number): pressure when the heart CONTRACTS. Pumps blood out. Diastolic (bottom number): pressure when the heart RELAXES. Refills. Normal: under 120/80. Elevated: 120-129 / under 80. Stage 1 hypertension: 130-139 / 80-89. Stage 2 hypertension: 140 plus / 90 plus. Hypertensive crisis: 180 plus / 120 plus - call 911.

1Predict your BP. Most teenagers run 105-120 / 60-80.

I understand systolic and diastolic

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 automatic BP cuff (upper-arm style, e.g., Omron), notebook. Optional: 1 stethoscope + manual sphygmomanometer for the bonus station. Quiet room. All cases anonymized; if a student has known BP issues, they observe instead.

How to Take Blood Pressure (Vital Signs)

Carefully watch this video to learn the tool before you use it. LevelUpRN walks through manual BP step by step. Watch this once before you cuff your partner.

The Hypothesis

1Predict your BP. Write systolic / diastolic.

The Build - Cuff Technique

2Take your blood pressure (cuff + stethoscope as shown in the video above). Open the Interpreter (above) and switch to BP mode. Enter your systolic / diastolic readings. The app classifies you as Normal, Elevated, Stage 1 Hypertension, Stage 2, or Hypertensive Crisis - same scale a real cardiologist uses.

3Take 4 readings: at rest, after 1 min of jumping jacks, 1 min after stopping, 5 min after stopping. Plot all 4 in the Interpreter. The app draws your recovery curve - that's a key fitness indicator.

4Sit quietly for 2 minutes before measurement. Feet flat on floor. Back supported.

5Place cuff on bare upper arm (not over clothes). The bottom edge of the cuff is 1 inch above the elbow crease.

6Rest your arm on the table at HEART LEVEL. Wrist relaxed, palm up.

7Press Start. Stay silent during the measurement. Record systolic / diastolic / pulse.

8Wait 2 minutes. Take a 2nd reading. Record. The two readings should be within 5 mmHg of each other.

9Now: stand up, do 30 seconds of standing in place arm circles. Sit. Take a 3rd reading immediately. Record.

10Wait 2 minutes. Take a 4th reading (recovery). Record.

Bonus: Capillary Refill Test

Why this test: when blood pressure is too low (shock), blood does not refill capillaries fast enough. EMTs check capillary refill on every patient as a quick perfusion screen. Normal: under 2 seconds. Slow: shock or dehydration. Combined with your BP reading, this is the EMT's '15-second snapshot' of cardiovascular status.

Capillary Refill Time Test - Normal vs Abnormal

Watch this short clinical demonstration. The technique is simple but the timing matters.

11Press firmly on the nail bed of your thumb for 5 seconds. The nail bed turns white as you push blood out.

12Release. Start counting (or use a stopwatch). How long does it take for pink color to return? Record the number of seconds. Try on 3 fingers, average them.

13If your reading is over 2 seconds, drink water and try again in 5 minutes. Many teenage refills are slightly slow simply due to dehydration. Record your final number next to your BP readings - the BP + cap refill pair tells a richer perfusion story than either alone.

If your BP reads stage 2 (140+/90+), do not panic. Cuffs misread on tense or anxious people. Take a 3rd quiet reading. Tell instructor.

I measured my BP at rest, after activity, and recovery

I followed proper cuff technique

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Reads the Trend

AI does pattern recognition only. A real cardiac patient sees a real cardiologist. Today's case is fictional.

1Open Gemini. Type: 'Fictional patient: 6 monthly BP readings: 118/76, 122/78, 128/82, 134/86, 138/88, 142/92. Describe the pattern. List 3 lifestyle factors that commonly produce this pattern.'

2Read the answer. Write the pattern in your notebook in one sentence.

3Ask: 'At what point in this trend should the patient see a doctor?' Read the answer.

This is exactly how 'patient portals' alert users to BP creeping up. The AI is a screening layer, not a diagnostician.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class BP Data

State	Class Avg Systolic	Class Avg Diastolic	Class Avg Pulse
Rest	fill	fill	fill
After activity	fill	fill	fill
2-min recovery	fill	fill	fill

1How much did systolic rise after activity? How fast did it recover?

Recovery rate is one of the most important predictors of cardiac fitness. The faster you return to baseline, the healthier the cardiovascular system.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Cardiology Tech

A Day in the Life of a Sonographer

Check out this video to see a real professional in this role. Cardiac sonographers run ultrasound studies just like the cardiac tech work you did today.

Cardiovascular Technologist: What they do: run EKGs, stress tests, BP monitoring; assist cardiologists in cath labs. Entry pathway: 1-2 year cert program at LA City College or East LA College. CA license required. Salary band (BLS LA 2024): entry 60,000 to 75,000. Mid-career 80,000 to 105,000. First step from where you sit today: tour the LAUSD ROP Health Career Pathways program - free, open to high schoolers.

Reflection

1Reflection 1: What did your BP recovery rate tell you about your fitness?

2Reflection 2: Why is silence required during BP measurement?

3Reflection 3: BP is silent. What other 'silent' health signals should you pay attention to?

I measured BP correctly

I understand the rest/active difference

Built by Woven Learning & Technology✎ Edit

Day 16: Heart and Lung Sounds - Auscultation

Diagnostic Tech Pathway - Auscultation Technique

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Yesterday you measured heart rate, breathing rate, and oxygen level. Today you measure the fourth vital: blood pressure.

How Do the Lungs Work?

Watch the video to set up today's thinking - this is the real-world story behind the lab. TED-Ed's animation of the respiratory system. The sounds you hear with the stethoscope come from this airflow.

Foundations - Two Sounds

For 200 years the stethoscope has been the most iconic tool in medicine. Today you learn what it does and how to use it - the same skill every doctor and nurse practices for years.

The Hook: In 1816 a French doctor named Rene Laennec felt awkward putting his ear directly on a young woman's chest to listen to her heart. He rolled up a piece of paper into a tube, put one end on her chest, his ear on the other - and discovered that her heart sounds came through LOUDER and CLEARER. The stethoscope was born. Same basic principle 200 years later.

S1 ('lub'): heart valves close as ventricles contract. Loud, low. S2 ('dub'): heart valves close as ventricles relax. Crisp, higher. Lung sounds: clear, soft 'whoosh' on inhalation and exhalation. Wheezing or crackling = abnormal. A full auscultation hits 5 spots on the chest and 6 on the back.

1Predict: how loud is your own heart?

I understand S1 and S2 sounds

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 stethoscope (acoustic), alcohol wipes, notebook. Quiet room. Each student listens to their OWN heart and lungs first. Plus for the lung capacity station: 2 round latex-free balloons per student (9 inch), measuring tape (cm), notebook. Optional: partner-on-partner listening over clothing.

How Breathing Works

Carefully watch this video to learn the tool before you use it. A focused look at the diaphragm. Watch this before lung-sound auscultation.

The Hypothesis

1Predict: how many distinct sounds will you hear from your heart in 10 seconds?

The Build - Listen

Cheat Sheet - 5 Stethoscope Positions: Map each spot on your chest. Listen for 30 seconds at each. 1. AORTIC: 2nd rib, RIGHT of sternum. Listen for S1/S2. 2. PULMONIC: 2nd rib, LEFT of sternum. S1/S2 again, slightly different. 3. ERB'S POINT: 3rd rib, LEFT of sternum. Best for hearing both valves clearly. 4. TRICUSPID: 4th rib, LEFT of sternum (lower). 5. MITRAL: 5th rib, just below the LEFT nipple line. THIS IS THE LOUDEST 'lub-dub' spot. Lung sounds: upper back at the shoulder blades. Listen for the soft 'whoosh' on inhalation. Wheezing or crackling = abnormal. Clean the diaphragm with alcohol between users. Real infection-control practice.

2Wipe the stethoscope diaphragm with alcohol. Put eartips in (angled forward toward your nose).

3Place the diaphragm on the LEFT side of your own chest, just below the collarbone. Listen for 30 seconds. Count the heart beats. How many sounds per beat?

4Move to the lower left chest (over the heart's apex). The 'lub-dub' should be loudest here. Record the loudest spot.

5Now move to your back. Place the diaphragm on the upper back. Take a deep breath. Listen for the 'whoosh' of air. Record the sound quality (clear, raspy, wheezy).

6Trade with your partner (over clothing). Listen to their heart at the apex. Compare to yours - faster, slower, louder, softer?

Station 4 - Lung Capacity (Balloon Spirometry)

Why this test: tidal volume (~500 mL) is what you breathe at rest. Vital capacity (3-5 L) is the MOST you can blow out after a deep breath. Hospitals measure this with a spirometer; we'll measure it with a balloon. Smokers, asthmatics, and athletes all have very different vital capacities.

Test Your Lung Capacity (Museum of Science DIY Experiment)

Watch this short DIY experiment first. The math is simple: balloon volume from circumference.

7Take 2 deep breaths to clear your lungs. Then take ONE big breath, as deep as you can.

8Blow ALL of that air into the balloon in a single breath. Pinch the balloon shut as soon as you stop blowing - no extra puffs.

9Have your partner measure the circumference around the WIDEST part of the balloon with the measuring tape. Record in cm.

10Calculate volume using the formula: V (mL) ≈ (4/3) × π × (C / (2π))^3 ≈ 0.0169 × C^3, where C is the circumference in cm. Or use this quick lookup: 30 cm ≈ 460 mL, 40 cm ≈ 1080 mL, 50 cm ≈ 2110 mL, 60 cm ≈ 3650 mL, 70 cm ≈ 5800 mL.

11Compare to typical ranges: average adult vital capacity is 3-5 L. Athletes can hit 5-7 L. Smokers and asthmatics often under 3 L. Try TWICE - your second attempt is usually better as you learn the technique.

12Build a table in your notebook: Method | BPM | Lung Capacity (mL) | Notes. Record YOUR data, then add 4 more pairs.

Wipe the diaphragm and eartips with alcohol between users. This is real infection-control practice.

I heard my own heart sounds

I heard my own lung sounds

I cleaned the stethoscope

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Stethoscopes Are Real

1Open Gemini. Type: 'How does an AI-powered stethoscope like Eko work? What can it detect that a human cannot?'

2Read the answer. Look for: heart murmurs, atrial fibrillation, low-amplitude sounds humans miss.

3Ask: 'What is the FDA approval status of AI stethoscopes?' This tells you if they are used in real clinics today.

AI cardiology is now real medicine. Eko is FDA-cleared. Patients in some clinics already get a 'second listen' from an AI that catches patterns the human did not.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Stethoscope vs Oximeter

Method	BPM	Notes
Stethoscope (you)	fill	fill
Oximeter (Day 14)	fill	fill
Difference	fill	fill

1Did the two methods give the same number? If different, why?

Both methods are valid. Oximeters use IR light reflection through your finger. Stethoscopes catch the actual mechanical sound. They almost always agree within 2 BPM.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Diagnostic Tech - Auscultation

Day in the Life of a Sonographer at Legacy Health

Check out this video to see a real professional in this role. Auscultation transitions into formal sonography work in real cardiology departments.

Diagnostic Medical Sonographer: What they do: use ultrasound, EKG, and other diagnostic tools to image and listen to the body. Entry pathway: 2-year associate degree program. Pasadena City College, Mt SAC, and Cypress College all offer it. Heavily impacted - apply early. Salary band (BLS LA 2024): entry 75,000 to 92,000. Mid-career 95,000 to 130,000. First step from where you sit today: shadow a sonographer at a Kaiser or USC Health clinic - LA TRIO can help arrange.

Reflection

1Reflection 1: What surprised you about hearing your own heart?

2Reflection 2: When have you noticed a sound that told you something was wrong?

3Reflection 3: How would you trust an AI to listen to a real patient?

I used a stethoscope correctly

I heard heart and lung sounds

Built by Woven Learning & Technology✎ Edit

Day 17: Biochemical Markers - Reading the Bloodwork

Lab Tech Pathway - Reading the Numbers

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Yesterday you used a stethoscope to listen to your heart and lungs. Today you read what's INSIDE your blood - the lab panels real doctors order on every patient.

How To Understand Your Blood Test Results

Watch this short video to see why today's topic matters before we dive in. A doctor walks through the panels you will read in today's lab - what each number means, and why one out-of-range value changes a clinical decision.

Foundations - The CBC and CMP

Vitals are real-time. Bloodwork is what your body has been doing for the last 3 months. Today you learn to read the most common lab tests in medicine.

The Hook: A 32-year-old woman with no symptoms goes for a routine annual blood test. Her white blood cell count is 47,000 (normal: 4,000 to 11,000). Her doctor calls her that night. She has chronic myeloid leukemia. Diagnosed in stage 1, CML has a 90 percent 5-year survival. Diagnosed in stage 4, it has 24 percent. One number on a routine test changed her life.

CBC = Complete Blood Count. Measures red cells, white cells, platelets, hemoglobin. CMP = Comprehensive Metabolic Panel. Measures glucose, kidney function, liver function, electrolytes. A CBC + CMP is run on EVERY hospital admission in America. Together they catch most major problems.

1Predict: what is your white blood cell count right now? (Normal: 4,000 to 11,000.)

I understand CBC and CMP

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 simulated blood-typing kit (Lab-Aids 'ABO/Rh Blood Typing' or Carolina Biological 'Simulated Blood Typing Activity', ~$30-40 per kit, covers 4-8 students). Each kit includes: 4 simulated blood samples (Patient A, B, C, D), anti-A serum, anti-B serum, anti-Rh serum (optional), 1 test card with labeled wells, toothpicks for stirring, gloves. Plus: notebook, pen. NO real blood is used. The 'blood' is a colored protein solution chosen specifically for school labs.

Today's task: act as a clinical blood-bank technician. Four 'patient samples' arrive on your bench labeled A, B, C, D. Using simulated antibody sera, you will determine each patient's ABO blood type. Then in a critical-care scenario, decide which patients can safely donate to which. This is the same chemistry every transfusion in the world depends on. Get it wrong - the patient's immune system attacks the donor blood and they die within minutes. Get it right - you save a life.

Why Blood Types Matter

Why Do Blood Types Matter? (TED-Ed)

Watch this 5-minute TED-Ed BEFORE the lab. The chemistry of why one wrong blood type can kill a patient in minutes - and the genius work of the 1900 lab tech who first mapped it out.

The Question

You have 4 'patient samples' on your bench. Patient C is in the ER bleeding out and needs an emergency transfusion. Which of patients A, B, or D can safely donate? You have 30 minutes to find out using ONLY ABO blood typing chemistry.

Your Hypothesis

1Before testing anything: predict which 'patient' will be each blood type. The four types are A, B, AB, and O - one each. Look at sample colors / appearance and write your guess for A, B, C, D in your notebook with reasoning. (Hint: appearance alone won't tell you - that is the LESSON. The chemistry is what reveals truth.)

How to Read Agglutination (Tutorial)

Blood Typing Lab Demo (Maria Berger)

Carefully watch this short demonstration BEFORE you test. Pay attention to how the demonstrator stirs each well with a toothpick and how they identify CLUMPING vs SMOOTH liquid. The clumps are tiny - hold the card up to a light to see them clearly.

The Build - Bench Lab

2Glove up. Set out the kit on your bench: 4 patient sample bottles (A, B, C, D), anti-A serum bottle, anti-B serum bottle, optional anti-Rh serum, test card with 12 wells (3 wells per patient: A / B / Rh), toothpicks.

3Patient A row: place 1 drop of Patient A's blood in EACH of 3 wells (anti-A column, anti-B column, anti-Rh column). Then add 1 drop of anti-A serum to the FIRST well, 1 drop of anti-B to the SECOND, 1 drop of anti-Rh to the THIRD.

4Stir each well with a SEPARATE toothpick (do not cross-contaminate). Stir for 15 seconds per well. Hold the card up to overhead light.

5Look at each well. CLUMPED = grainy, lumpy texture (positive reaction). SMOOTH = uniform liquid (negative reaction). Record + or - for each well in your notebook.

6Determine Patient A's blood type from the table: anti-A clumps + anti-B clumps = TYPE AB anti-A clumps + anti-B smooth = TYPE A anti-A smooth + anti-B clumps = TYPE B anti-A smooth + anti-B smooth = TYPE O anti-Rh clumps = + (positive) anti-Rh smooth = - (negative)

7Repeat for Patients B, C, D using fresh wells (or wipe and reuse the card if your kit specifies).

Copy this table into your Woven notebook BEFORE class data collection. Sketch the column headers neatly on a fresh page. As pairs report data, fill in your notebook copy AND watch the teacher fill the same table on the whiteboard / slide.

Patient Blood Type Results

Patient	anti-A	anti-B	anti-Rh	Blood Type	Donate to?	Receive from?
A
B
C
D

Critical Care Decision

8Patient C (the bleeder in the ER) needs immediate transfusion. Look at your blood-type results. Which of A, B, or D can safely donate to C? Which CANNOT (would kill the patient)? Justify your answer using the agglutination data from your test, NOT your hypothesis from earlier.

9Bonus: in the 1940s, blood banks figured out that ONE patient type can donate to anyone (universal donor) and ONE patient type can receive from anyone (universal recipient). Look at your results - identify both. Why is the universal donor / recipient pair the most valuable in a hospital blood bank?

10Compare your initial hypothesis (sample-appearance guess) to your tested blood-type results. Were you right? The lesson: appearance tells you NOTHING about blood compatibility. The chemistry is what saves lives.

Treat this like a real lab. Gloves on. No eating or drinking at the bench. Used toothpicks + gloves go in the bench trash can. Wipe up spills immediately. The simulated blood is non-toxic but stains clothes.

I read 3 fictional lab panels

I categorized urgency

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Triages the Panel

1Open Gemini. Paste Patient C's labs (the most concerning one). Type: 'These are fictional labs for a teaching exercise. Identify the pattern. Do not diagnose. List urgency level: routine, follow-up, or emergent.'

2Read the AI's pattern. Did it agree with your assessment?

3Ask: 'What single follow-up test would best clarify this pattern?' Note the answer.

This exact pattern (high WBC + low Hgb + high glucose + high creatinine) is consistent with several conditions including diabetic ketoacidosis with infection. A real ER doctor would order more tests.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Triage

Patient	Your Verdict	Class Verdict	AI Verdict
A	fill	fill	fill
B	fill	fill	fill
C	fill	fill	fill

1Where did the class disagree most? Where did the AI disagree with the class?

Real ER triage relies on multiple readers - that is why hospitals require physician sign-off plus charge-nurse review for major decisions.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Lab Tech - Reading the Numbers

Day in the Life of a Medical Laboratory Technician

Check out this video to see a real professional in this role. Medical lab techs read the same panels you triaged today, all day, every shift.

Medical Laboratory Technician (MLT): What they do: run the lab tests that produce CBCs, CMPs, urinalysis, and dozens more. Behind every doctor visit there is a lab tech making the data real. Entry pathway: 2-year associate degree at LA City College or Cypress College. CA license required. Salary band (BLS LA 2024): entry 60,000 to 75,000. Mid-career 80,000 to 105,000. First step from where you sit today: take AP Biology or college Biology in junior year, apply to a hospital phlebotomy internship in summer.

Reflection

1Reflection 1: What surprised you about reading lab data?

2Reflection 2: When have numbers told you something hidden?

3Reflection 3: How is reading a lab panel like reading drone telemetry?

I read 3 fictional panels

I categorized urgency

Built by Woven Learning & Technology✎ Edit

Day 18: AI Diagnostic-Support - Pattern Recognition Lab

Clinical Research Pathway - AI as Critic

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Yesterday you read fictional bloodwork. Today you bring AI in to triage harder cases - and check whether you trust its calls.

AI in High-Stakes Decisions

Carefully view this video to ground today's hook in a real example. Same Coded Bias clip. Today we apply the same skepticism to medical AI.

Foundations - Three AI Roles in Medicine

AI in medicine is one of the most exciting and most controversial areas in healthcare. Today you see what it can and cannot do - and write your own opinion on where the line should be.

The Hook: In 2023, a Stanford study put a Google AI in a clinic. The AI looked at chest X-rays alongside radiologists. The result? AI alone: 88 percent accurate. Radiologist alone: 91 percent accurate. Radiologist + AI together: 96 percent accurate. The AI was not better. The COMBINATION was better. That is the future of medicine.

Role 1: SCREENING. AI looks at a million scans and flags the 5 percent that need a human's attention. Role 2: SECOND OPINION. AI checks every doctor's diagnosis and flags potential errors. Role 3: ASSISTANT. AI drafts notes, summarizes, transcribes - so doctors spend more time with patients. Role 4 (controversial): primary diagnosis. Currently illegal in California without physician sign-off.

1Predict: which of those roles do you think is most useful?

I understand the 4 roles of AI in medicine

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 wooden ruler (30 cm, with cm markings clearly visible), 1 reflex hammer OR rolled-up notebook spine, phone with flashlight (each pair has access), notebook, laptop with Gemini or Claude open. The 5 fictional patient case cards from Days 14-17 are still available in Phase 3 for the AI-comparison portion. All cases are anonymized fictional teaching examples.

The Question

What is the difference between a REFLEX (autonomic, ~50 ms, no brain) and a REACTION (cognitive, ~200 ms, brain involved)? Today you measure all three classic clinical tests on yourself - ruler drop (reaction time), patellar (knee jerk reflex), and pupillary (light reflex). Then in Phase 3 you compare your numbers to clinical normal ranges using AI.

Your Hypothesis

1Before you test anything, predict in your notebook: (1) what is YOUR reaction time in milliseconds? (Adult average ~200 ms.) (2) Will your patellar reflex be larger, smaller, or the same as your reaction time? (3) Which feels faster - knee jerk or pupil constriction? Write a number + a one-sentence reason.

Station 1 - Reaction Time (Ruler Drop Test)

Why this test: a reaction involves your eye seeing the ruler drop, your brain deciding 'CATCH IT', and your hand muscles closing - 3 separate signals through the brain. The total time is your reaction time. Athletes train this; gamers obsess over it; pilots are tested for it.

Reaction Time: The Ruler Drop Test (Kids Fun Science)

Carefully watch this short demonstration BEFORE you test. Pay attention to where partner's hand sits at the bottom of the ruler - that hand position is the start point for your distance measurement.

2Partner holds the ruler vertically, with the 0 cm end at the bottom. Your thumb and forefinger are positioned at the 0 cm mark with a slight gap (about 2-3 cm wide), ready to pinch.

3Partner releases the ruler WITHOUT saying 'go' - random timing. You catch it as fast as you can. Read the cm mark at your top finger. That is your DROP DISTANCE in cm.

4Run 5 trials. Throw out the highest and lowest. Average the middle 3. Record your average drop distance.

5Convert distance to reaction time: t (ms) = sqrt(2 × d / 9.81) × 1000, where d is in METERS (drop_cm / 100). Or use this lookup table: 5 cm = 101 ms, 10 cm = 143 ms, 15 cm = 175 ms, 20 cm = 202 ms, 25 cm = 226 ms, 30 cm = 247 ms. Record your reaction time in milliseconds.

Station 2 - Patellar Reflex (Knee Jerk)

Why this test: the patellar reflex is one of the FASTEST reflexes in your body - signal travels only to the spinal cord and back, NEVER reaching your brain. This is why it happens in 30-50 ms. Doctors check it because absent reflex can signal nerve damage; exaggerated reflex can signal upper motor neuron problems.

Patella Tendon Reflex Demonstration

Watch this short physical-therapy demonstration BEFORE you try it. Same technique a physician uses during a routine physical.

6Sit on a chair with one leg crossed over the other so your foot dangles freely. Relax the leg completely - do NOT tense the quadriceps.

7Partner taps firmly but not too hard with the reflex hammer (or notebook spine) JUST BELOW your kneecap, on the patellar tendon (the soft spot between knee and shinbone).

8Observe: does your foot kick forward involuntarily? How big is the response? Record on a 0-3 scale: 0 = no response, 1 = small kick, 2 = normal kick, 3 = exaggerated. Try 3 taps to confirm.

9KEY OBSERVATION: how fast did the kick happen after the tap? Patellar reflex completes in ~50 ms - way faster than your reaction time (~200 ms). Why? The signal does not go to your brain. It bounces off the spinal cord.

Station 3 - Pupillary Light Reflex (PERRLA)

Why this test: pupils react in ~200 ms to light. The reflex is automatic but routes through the brainstem (not the spinal cord like patellar). EMTs check this every shift because absent or sluggish pupil response can signal a brain bleed, drug overdose, or severe shock.

Pupillary Light Reflex Assessment (RegisteredNurseRN)

Carefully watch this 1-minute clinical demonstration BEFORE you try it. Same technique an ER nurse uses on every patient.

10Dim the room slightly (close blinds or shade your test area). Partner stands facing you and looks at a fixed point past your ear (NOT at the light).

11You shine your phone flashlight INTO their right eye for 2 seconds (cast the beam from the side, not directly into the pupil at close range). Observe BOTH pupils - they should constrict at the same time (consensual response).

12Repeat on the left eye. Then swap roles. Record on a 0-3 scale: 0 = no response (concerning), 1 = sluggish, 2 = normal direct + consensual, 3 = brisk.

Compare Reflex vs Reaction

13Look at your 3 numbers: ruler-drop reaction time (~200 ms), patellar reflex speed (~50 ms by observation), pupillary reflex speed (~200 ms). Write a 1-sentence answer in your notebook: 'A reflex is __, a reaction is __, and the difference matters because __.'

Copy this table into your Woven notebook BEFORE class data collection starts. Sketch the column headers neatly on a fresh page. As pairs report data, fill in your notebook copy AND watch the teacher fill the same table on the whiteboard / slide.

Class Reflex + Reaction Data

Pair	Ruler Drop (cm)	Reaction Time (ms)	Patellar (0-3)	Pupillary (0-3)	Notes
You
Pair 2
Pair 3
Pair 4

14After class data is in: who has the fastest reaction time? Who has the most exaggerated patellar reflex? In your notebook, write 1 sentence about what factors might explain the spread (caffeine? sleep? gaming hours? sport?).

These are fictional. NEVER use AI to triage a real person. Real triage is a licensed clinical judgment.

I triaged 5 fictional cases

I compared with AI

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

Why this app: in Phase 2 you measured your own reflexes and reaction times. Now you have TWO ways to use AI as a clinical analyst. First, feed YOUR real numbers to Gemini and ask for clinical interpretation (is your reaction time normal? what does it tell you?). Second, run the Diagnostic Pattern Matcher (below) on 5 fictional patient cases - the app shows your call vs the AI's call so you can see where you and the AI agree or diverge. Use both.

1FIRST: open Gemini in a separate tab. Type: 'I just ran 3 clinical reflex tests on myself. My ruler-drop reaction time was [X] ms. My patellar reflex was [Y on 0-3 scale]. My pupillary reflex was [Z on 0-3 scale]. What is the clinical interpretation? Is anything outside the normal range? What factors could affect these numbers?' Read the AI's response. Compare to what YOU would say about your own data. Write the AI's clinical interpretation in your notebook.

2Open the AI-compare view (above). Run all 5 cases again - the app feeds the same case to the AI and shows AI's call next to your call. Note where you and the AI agreed. Note where you disagreed.

3Pick the case where you and the AI disagreed MOST. Read the AI's reasoning carefully. Write in your notebook: (1) is the AI right, are YOU right, or is neither? (2) what would you ask the patient or test for to break the tie?

Find the AI's Mistake

4Pick the case where you and the AI disagreed most. Re-read the AI's reasoning.

5Write in your notebook: WHO is right - you, the AI, or neither? Why?

6Ask AI: 'What additional information would change your triage?' Read the answer.

This is the most important AI literacy skill: NOT taking the AI's word for it. The AI is a tool. You are the user. The user is responsible.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Agreement Rate

Case	Class Avg	AI Call	Match (%)
1	fill	fill	fill
2	fill	fill	fill
3	fill	fill	fill
4	fill	fill	fill
5	fill	fill	fill

1What was the overall agreement rate? Which case had the biggest disagreement? Why?

In real radiology AI studies, the agreement rate between AI and humans is around 85 percent. The 15 percent of disagreements is where the most learning happens for both AI and humans.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Clinical Research - AI as Critic

Medical Scientist Career Video

Watch this video to picture yourself in this career 5 to 10 years from now. Clinical researchers test new tools (drugs, devices, AI) the way you tested AI critique today.

Clinical Research Coordinator / AI in Healthcare: What they do: design studies that test whether new tools (drugs, devices, AI) actually work. The bridge between research labs and real patients. Entry pathway: BS in Biology, Public Health, or Health Sciences. Cal State LA's Public Health BS is direct. UCLA, USC, and UCSF all have clinical research masters programs. Salary band (BLS LA 2024): entry 75,000 to 95,000. Mid-career 100,000 to 140,000. First step from where you sit today: apply to UCLA's Summer Research Program for High School Students.

Reflection

1Reflection 1: When did the AI surprise you - either by being right or by being wrong?

2Reflection 2: Where do you draw the line for 'AI helped' vs 'AI replaced human judgment'?

3Reflection 3: If you were a doctor, would you want AI in your exam room? Why or why not?

Week 4 Complete - med students: Five days of vital signs, blood pressure, auscultation, bloodwork, and AI diagnostic-support. You measured your own body and triaged fictional patients with real numbers. Stand up. Read your most surprising vital reading out loud. You finished Week 4. Group photo. Sign your notebook. See you next session.

I completed Week 4

I formed an opinion on AI in medicine

Built by Woven Learning & Technology✎ Edit

Day 19: Energy Conversion Basics - Photons to Electrons

Environmental Engineer Pathway - Solar Fundamentals

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

How Do Solar Panels Work?

Carefully view this video to ground today's hook in a real example. TED-Ed's animation of the photoelectric effect. The voltage you measure today is photons becoming electrons.

Foundations - Efficiency

From medicine to the planet. This week is about energy - how we capture it from sunlight and wind, and how we make those captures more efficient. The science is real, the math is solvable, and the careers are growing fastest of any field in California.

The Hook: In 2024 California generated 20 percent of its electricity from solar - more than any other state in the US, more than any country in Europe except Germany. The Mojave Desert solar farm alone powers 1.2 million homes. Every one of those solar panels does the same thing the small solar cell on your bench does: turn photons into electrons. Today you measure exactly how efficiently.

Efficiency = Useful Energy Out / Total Energy In. A top-grade silicon solar cell: 22 percent efficient. A classroom toy solar cell: 5-12 percent. The theoretical maximum (Shockley-Queisser limit): 33 percent. The sun's energy at Earth's surface: about 1000 watts per square meter at noon. The rest of the energy becomes HEAT. That is why solar panels get warm.

1Predict: how much voltage will a small solar cell produce in direct light?

I understand efficiency

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 small solar cell (1V to 6V output), 1 multimeter, 1 desk lamp (or sunlight access), measuring tape, notebook. Plus for the prototype build station: 1 small DC motor (3V hobby motor, ~$2 each), 1 sheet cardstock for the propeller, 2 alligator-clip leads OR jumper wires (likely already used for the multimeter), tape, scissors.

Energy 101: Wind Power (US Dept of Energy)

Check out this video to learn the steps before you do them yourself. This week is solar AND wind - watch the DOE's overview of how a wind turbine converts moving air into electricity, then on Day 21 you build your own.

The Hypothesis

1Predict the voltage at: 10 cm from lamp, 30 cm, 50 cm, 1 m. Write 4 numbers.

The Build - Solar Cell Characterization

Cheat Sheet - Inverse Square Law: Light intensity drops as 1 over distance squared. Intensity at distance d = Intensity at d=1 divided by (d squared) Worked example for your lab: At 10 cm: voltage X (call this 100%) At 20 cm: voltage = X / (2 squared) = X / 4 = 25% At 30 cm: voltage = X / (3 squared) = X / 9 = ~11% At 100 cm: voltage = X / (10 squared) = X / 100 = ~1% The law applies to: sunlight, gravity, sound, Wi-Fi signal. Knowing it is foundational physics. After your lab, see how close your data fits this curve.

2Set multimeter to DC voltage (V). Connect to solar cell leads.

3Hold cell 10 cm from lamp. Wait 10 seconds. Record voltage.

4Move to 30 cm. Record. Then 50 cm. Then 100 cm.

5Now switch multimeter to current (mA). Repeat the 4 distances.

6Calculate POWER for each distance: P = V x I (volts times amps). Use mA divided by 1000 = A.

7Build a table: Distance, Voltage, Current, Power.

The Build - Solar-Powered Fan Prototype

Why build this: a multimeter shows you a NUMBER. A spinning propeller shows you ENERGY in motion - the same energy your data table just measured, but now you can see and hear it. Engineers prototype because numbers alone do not tell you whether a design works in the real world. This 10-minute build proves your solar cell delivers usable mechanical power, not just a voltage reading.

How to Make a Solar-Powered Fan (DIY)

Watch this short tutorial first. Yours will be simpler - just cell, motor, and a paper propeller taped to the motor shaft.

8Cut a propeller from cardstock: 6 cm x 1.5 cm rectangle, twist gently so each end angles ~30 degrees in opposite directions. Tape the center to the motor shaft so the propeller spins freely without hitting the motor body.

9Connect the motor terminals to the solar cell leads using the alligator clip leads (red to red, black to black is conventional but either way works for DC).

10Hold the assembled prototype 10 cm from the lamp, propeller facing UP and free of obstruction. Does it spin? If not - check connections, try moving closer to the lamp, or check that the propeller blade angle is creating thrust (not flat).

11Once spinning: move the prototype to the SAME 4 distances you tested earlier (10, 30, 50, 100 cm). Time how many full rotations the propeller does in 10 seconds at each distance. Record.

12Compare your fan-rotation data to your voltage / current / power data from earlier. Does the propeller speed track with the power calculation? At what distance does the propeller stop spinning entirely? That is your minimum-viable-power threshold for THIS motor.

13Bonus: place a piece of paper between the cell and the lamp. Voltage now? Calculate the percent drop.

Lamps get HOT. Do not touch the bulb. Keep the cell at least 5 cm away to prevent thermal damage.

I measured voltage and current at 4 distances

I calculated power

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Finds the Law

1Open Gemini. Type: 'My solar cell produces these voltages at these distances from a lamp: [paste your 4 data points]. What law of physics describes this relationship?'

2Read the answer. The AI should mention the INVERSE SQUARE LAW: intensity drops as 1/r squared.

3Ask: 'Predict voltage at 75 cm.' Test the prediction with your multimeter.

This same law (inverse square) governs sunlight, gravity, sound, and Wi-Fi signal. Knowing it is foundational physics.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Solar Data

Distance (cm)	Your Power (mW)	Class Avg (mW)
10	fill	fill
30	fill	fill
50	fill	fill
100	fill	fill

1Plot the curve. Distance on X, Power on Y. What shape is it?

This curve is identical to what real solar engineers see - just scaled up. A 100x larger cell at 100x the distance gives the same proportional power.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Environmental Engineer - Solar Fundamentals

A Day in the Life of a Solar Power Technician

Check out this video to see a real professional in this role. Solar power technicians install and tune the same kind of cells you measured today.

Solar PV Engineer: What they do: design, install, and optimize solar power systems for utilities, businesses, and homes. Entry pathway: BS Electrical Engineering or Mechanical Engineering with renewable-energy focus. Cal State LA's EE program with sustainability concentration. ALSO: 1-year SOLAR INSTALLER certification at Mt SAC College for fast entry to fieldwork. Salary band (BLS LA 2024): entry installer 60,000 to 75,000. Engineer entry 90,000 to 110,000. Senior engineer 140,000 to 180,000. First step from where you sit today: get NABCEP entry-level certification (online, low cost) to flag yourself for solar internships.

Reflection

1Reflection 1: What surprised you about how power dropped with distance?

2Reflection 2: Where in your home are there 'energy losses' you never thought about?

3Reflection 3: If you ran LA's energy grid, what would your 10-year solar plan look like?

Spotlight: Dr. Shirley Ann Jackson: Dr. Jackson (1946 to present) was the first Black woman to earn a PhD from MIT. As a theoretical physicist at Bell Labs in the 1970s and 1980s, her research enabled portable fax machines, touch-tone telephones, fiber-optic cables, solar cells, and caller ID. She later led the US Nuclear Regulatory Commission and served as President of Rensselaer Polytechnic Institute for 23 years. Her physics is in your phone. Her oversight is in every nuclear plant in the country. The energy work you do today connects to the systems she designed and regulates.

I measured solar power

I derived the inverse-square law

Built by Woven Learning & Technology✎ Edit

Day 20: Light Filters and Solar Efficiency

Environmental Engineer Pathway - Spectral Response

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Solar Panels Recap

Watch this short recap to bring back the key idea. Same explainer, this time as we test how light filters change output.

Foundations - Photons Have Energy

Yesterday you measured how distance affects solar output. Today you find out something stranger: COLOR matters. A solar cell does not respond equally to red, green, and blue light.

The Hook: The sun emits a rainbow of light - infrared, red, orange, yellow, green, blue, violet, UV. A silicon solar cell catches them at different efficiencies. RED light: 80 percent of available energy captured. BLUE light: 30 percent. UV: nearly 0. This is called the spectral response curve. It is why solar farm engineers design panels with multiple cell types - so each color of light has a cell tuned to it.

Photons of different colors carry different energies. Blue photons: high energy. Red photons: low energy. A silicon cell has a specific 'band gap' - it can only absorb photons above a certain energy. Most red photons exceed it just barely. Blue photons exceed it by a lot, and the extra energy is wasted as heat. The sweet spot is RED for silicon cells.

1Predict: which color filter (red, green, blue) will let your solar cell produce the most voltage?

I understand spectral response

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: solar cell + multimeter + lamp from Day 19, color filter sheets (red, green, blue, yellow), tape, notebook.

The Hypothesis

1Predict the order from highest voltage to lowest: white (no filter), red, green, blue, yellow.

The Build - Spectral Test

2Open the Solar Filter Plotter (above). It plots PREDICTED voltage for white, red, green, blue, yellow filters based on silicon's spectral response. Enter your panel's no-filter voltage as the reference. The app fills in predictions for the colored filters.

3Run your real trials - measure voltage with each colored filter. Plot real values next to the prediction line. Where does reality match the model? Where does it diverge? Write 1-2 sentences explaining the gap.

4Set up cell at 30 cm from lamp. Measure baseline voltage with NO filter. Record.

5Tape RED filter over the cell. Re-measure. Record.

6Replace with GREEN filter. Record.

7BLUE filter. Record.

8YELLOW filter. Record.

9Calculate the percent drop from baseline for each filter.

10Build a table: Filter, Voltage, Percent of Baseline. Rank from highest to lowest.

Make sure the filter covers the full cell. Partial coverage gives mixed results.

I tested 4 colors

I ranked spectral response

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Predicts the Best Color

1Open Gemini. Type: 'For a silicon solar cell, which color of light produces the highest voltage: red, green, blue, or yellow? Explain why.'

2Read the answer. Compare to your data - did the AI's prediction match?

3Ask: 'How would multi-junction solar cells solve this color problem?' Read the answer (it should mention stacking different cell types).

NASA's Mars rovers use multi-junction cells - they hit 35 percent efficiency in space, more than any silicon cell on Earth. The trick is matching cell type to wavelength.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Spectral Data

Filter	Class Avg Voltage	% of Baseline	AI Predicted Rank
None	fill	100%	-
Red	fill	fill	fill
Green	fill	fill	fill
Blue	fill	fill	fill
Yellow	fill	fill	fill

1Did the class data match the AI prediction? Where was it different?

This curve is exactly what spectral-response engineers measure on every new cell type. Your data IS the same data they collect.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Environmental Engineer - Spectral Response

A Day in the Life of a Renewable Energy Developer

Check out this video to see a real professional in this role. Renewable energy developers connect solar/wind to real grid demand.

Renewable Energy Engineer: What they do: design and optimize solar arrays, wind farms, and battery storage for utilities and corporations. Entry pathway: BS Electrical or Mechanical Engineering. Cal State LA + Cal Poly Pomona transfer track is local. UCLA and USC also offer renewables tracks. Salary band (BLS LA 2024): entry 90,000 to 110,000. Mid-career 130,000 to 175,000. First step from where you sit today: apply to Sunrun's High School Internship Program (LA-based, paid).

Reflection

1Reflection 1: What was surprising about how color affected the cell?

2Reflection 2: When in your life does the same input produce different results based on a hidden factor?

3Reflection 3: If you designed a solar farm in LA, what colors of light would you optimize for at sunset?

I tested 4 colors

I beat or matched the AI's prediction

Built by Woven Learning & Technology✎ Edit

Day 21: Wind Turbine Prototypes

Environmental Engineer Pathway - Mechanical Energy Capture

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Yesterday you tested how light filters change solar voltage. Today you switch to wind and build a real turbine prototype.

How Wind Turbines Generate Power (DOE)

Watch this short video to see why today's topic matters before we dive in. Same DOE Energy 101 explainer - this time pay attention to blade pitch and number of blades, because those are the variables you control today.

Foundations - Blade Variables

From sunlight to wind. The Tehachapi wind farm in LA County has 5,000 turbines and powers half a million homes. Today you build a baby version and find out what makes a turbine efficient.

The Hook: The Tehachapi wind farm is 60 miles north of LA. Its 5,000 turbines generate 1,300 megawatts - enough for 500,000 homes. Each turbine blade is 200 feet long and weighs 16 tons. Despite the size, the physics that makes them spin is the same as the toy turbine in front of you: incoming air pushes a tilted blade. Efficient blades capture more of the wind's energy. Inefficient blades let it pass through.

Three variables that drive blade efficiency: 1. Number of blades (more = more drag, fewer = less capture) 2. Pitch angle (the tilt - too flat catches no wind, too steep stalls) 3. Blade length (longer = more capture but harder to spin) Real turbines: 3 blades, 4-6 degree pitch, as long as engineering allows.

1Predict: which blade design will produce the most voltage from a fan: 2 blades, 3 blades, or 5 blades?

I can name 3 blade variables

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: 1 small DC generator (or hobby motor used as generator), 1 hub or rotor, cardstock, scissors, tape, ruler, kitchen scale, multimeter, 1 small fan (set up at instructor station). Notebook.

The Hypothesis

1Sketch your blade design in your notebook. Label number of blades, length, and pitch angle. Predict voltage.

The Build - Wind Prototype

2Open the Wind Turbine Grapher (above). Enter your blade design: number of blades (3? 5? 8?), blade length, and pitch angle. The app predicts voltage at low, medium, and high wind speeds.

3Build your real prototype to those specs. Test in front of the fan at all 3 speeds. Plot your REAL voltage on the same graph as the PREDICTION. The grapher draws both lines - is your design above or below the model? Why?

4Cut 3 identical blades from cardstock: 10 cm long, 3 cm wide. Tape each at a 30-degree angle to the hub.

5Connect generator to multimeter. Set to DC volts.

6Hold prototype 30 cm in front of the fan. Turn fan on. Measure peak voltage. Record.

7Iterate: try 2 blades vs 5 blades. Same length, same angle. Measure each. Record.

8Iterate: change pitch to 15 degrees, then 45 degrees. Measure each. Record.

9Find your best combination of blade count and pitch. Record the winning numbers.

10Build a table: Trial, Blades, Pitch, Voltage.

Wind turbines spin FAST - keep fingers clear of blades when fan is on. Stop the fan before adjusting blades.

I tested at least 3 blade configurations

I found my best combination

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Optimizes the Blade

1Open Gemini. Type: 'For a small wind turbine in a 5 mph wind, what is the optimal blade pitch angle and number of blades for maximum power output?'

2Read the answer. Compare to your best configuration.

3Ask: 'Why do industrial wind turbines have only 3 blades when more would seem to capture more wind?' Read the answer.

The answer: more blades = more drag = slower spin. Three is the sweet spot for nearly all wind speeds. Engineering iterations over 50 years converged on this answer.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Wind Data

Pair	Blades	Pitch	Best Voltage (V)
Yours	fill	fill	fill
Class Best	fill	fill	fill
AI Suggestion	fill	fill	fill

1What configuration won? Did anyone beat the AI's suggestion?

In real engineering, students sometimes beat textbook answers because real wind is different from theoretical wind. The Tehachapi wind farm uses pitch sensors that adjust blade angle 60 times a second.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Environmental Engineer - Mechanical Energy Capture

Energy 101: Wind Power (Career Connection)

Check out this video to see a real professional in this role. Same DOE explainer you watched in Phase 1 - wind energy techs and engineers are the people building these full-size turbines.

Wind Energy Engineer / Wind Turbine Tech: What they do: design wind farms, maintain turbines, optimize blade design. Entry pathway: TWO routes. Engineer route: BS Mechanical or Aerospace at Cal State LA. Tech route: 1-year cert at Mt SAC College in Wind Energy Technology - hiring at 60-80k starting. Salary band (BLS LA 2024): wind tech entry 60,000 to 80,000. Engineer entry 95,000 to 115,000. Both grow fast. First step from where you sit today: tour Tehachapi Wind Farm (free public tours April-October).

Reflection

1Reflection 1: What was the most counterintuitive thing about blade design?

2Reflection 2: Where in your life does adding 'more' actually make something worse?

3Reflection 3: How would you design a wind turbine for an LA highway median?

I built a wind turbine

I optimized it

Built by Woven Learning & Technology✎ Edit

Day 22: Generative AI for Sustainable Infrastructure

Environmental Engineer Pathway - AI as Design Partner

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Yesterday you built a wind turbine. Today you bring AI in to design future-ready sustainable buildings - and find their flaws.

Gemini Canvas Tutorial (Vibe Coding)

Carefully view this video to ground today's hook in a real example. Mark Kashef's tutorial on Gemini Canvas (the tool you use today) - shows how anyone can vibe-code a design without writing traditional code.

Foundations - The Two Roles

You measured solar. You built a turbine. Today you use AI to DESIGN something at scale - a sustainable building or transit hub. The AI is your sketch artist; you are the engineer who decides what is real.

The Hook: In 2024, an LA architecture firm used Midjourney and DALL-E to generate over 1,000 sketches for a single sustainable transit hub project. The lead architect picked 12 to develop. Of those, 3 became real buildings. Generative AI is now a design tool, not a toy. Used right, it accelerates engineers by 10x. Used wrong, it generates beautiful but un-buildable nonsense.

AI as ARTIST: generates a visual concept based on words. AI as ENGINEER: would ensure the concept is physically buildable. Generative AI today is a great artist and a TERRIBLE engineer. It will draw a 100-story building with a foundation the size of a closet. Your job is to spot the impossible.

1Predict: when you ask AI to design a 'solar-powered LA bus station,' what 3 features will it include?

I understand AI's two roles

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: laptop with Gemini Image Gen or Canva AI access, notebook, pencil. Optional: printer to print the best designs.

Today's task with Gemini Image Gen: generate sustainable building designs (a 5-story apartment, a school, a transit hub) and watch the AI's first attempts. Tomorrow you flip the lens and use AI to critique YOUR designs. Code Here - Gemini Image Gen: Web app: https://gemini.google.com (sign in with your school Google account). Click the 'Imagine' tool or paste your prompt directly with 'Generate an image of...'. Wait 5-15 seconds for the rendering. If your prompt did not produce what you wanted, refine the prompt; do not blame the model on the first try.

How to Use Gemini Gems (Tutorial for Beginners)

Carefully watch this video to learn the tool before you use it. Simpletivity walks through how to build and use a custom Gem in Gemini - the EXACT skill you'll need for today's design generation.

Gemini Canvas Tutorial (Walkthrough)

Carefully watch this tutorial to learn the steps before you try them. Quick walkthrough of Gemini Canvas - how to prompt, iterate, and refine a design.

The Hypothesis

1Sketch the kind of sustainable building you would design (apartments, school, transit hub). Predict 3 features the AI will get RIGHT and 3 it will get WRONG.

The Build - Generate and Critique

2Open Gemini Image Gen or Canva AI. Prompt: 'Design a 5-story sustainable apartment building in Los Angeles with solar panels, wind turbines, and rainwater capture. Architectural rendering.'

3Save the image. In your notebook, list 3 features you can identify in the rendering.

4Critique: based on what you learned in days 19-21, find 3 ENGINEERING MISTAKES. (Examples: wind turbines too small to power building, solar at the wrong angle for LA latitude, no shade in the rendering.)

5Iterate: prompt the AI to FIX one of the mistakes. Generate again. Did it actually fix it?

6Repeat: 3 total iterations. Save each version.

7Pick your favorite final version. Sketch it in your notebook with annotations explaining what is real and what is wishful thinking.

AI image gen makes everything look possible. Your engineering mind is the filter. Without it, the rendering is fantasy.

I generated and critiqued 3 designs

I identified 3 engineering mistakes

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

Get a Second AI Opinion

1Switch to text Gemini. Describe your AI-generated building in 3 sentences. Ask: 'Is this design physically buildable in Los Angeles? List 3 potential problems.'

2Read the response. Did the text AI catch problems your eyes did not?

3Compare: which is the better critic - your engineering mind, or the text AI?

This is how real architecture firms work today. Image AI generates concepts. Text AI does feasibility checks. Human engineers make the final call. Three layers.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Design Showcase

1Pin your final design at the front. Walk through the gallery. Vote: most BUILDABLE, most BEAUTIFUL, most AMBITIOUS.

Design Award	Pair	Vote Count
Most Buildable	fill	fill
Most Beautiful	fill	fill
Most Ambitious	fill	fill

In real architecture awards, these are 3 different categories. Buildability and ambition often pull in opposite directions - the engineer's job is to find the middle.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Environmental Engineer - AI as Design Partner

A Day in the Life of a Researcher

Watch this video to see what a real day in this career actually looks like. Architects and civil engineers using AI tools work like the researchers in this clip.

Sustainable Architecture / Civil Engineer with AI Tools: What they do: design buildings, transit, and infrastructure that meet LEED Platinum sustainability standards. Use AI tools to iterate faster. Entry pathway: BS Civil Engineering or Architecture. Cal State LA + transfer to USC or Cal Poly. Add: get an internship at Gensler, AECOM, or Skidmore Owings & Merrill - all LA-based. Salary band (BLS LA 2024): entry 75,000 to 95,000. Mid-career 110,000 to 150,000. Senior 175,000 plus. First step from where you sit today: enter the Skyline AEC High School Design Challenge (annual, free).

Reflection

1Reflection 1: What was beautiful but unbuildable in your design?

2Reflection 2: When did the AI surprise you - good or bad?

3Reflection 3: If you could build ONE structure in LA, what would it be?

I generated 3 AI designs

I critiqued each

Built by Woven Learning & Technology✎ Edit

Day 23: AI Flaw Detection - Critique Your Prototype

Environmental Engineer Pathway - AI as Critic

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Today's goal: get an AI to find 5 engineering flaws in your favorite project from this week, then decide which critiques are real and which are AI hallucination.

Gemini Canvas (Recap)

Watch this short recap to bring back the key idea. Same Gemini Canvas tutorial from Day 22 - this time the lens shifts from 'AI as designer' to 'AI as critic'.

Foundations - Flaw Categories

Yesterday AI was your sketch artist. Today AI is your critic. You go back to your week's work - solar setup, wind turbine, building design - and feed it to AI for FLAW DETECTION.

The Hook: Elon Musk has a rule at SpaceX: 'The best part is no part. The best process is no process.' His engineers spend more time DELETING things than adding them. Flaw detection is the most underrated engineering skill on earth. AI is a powerful tool here - it does not get attached to your work.

Categories of engineering flaws AI can usually catch: 1. Physics violations (impossible math) 2. Code/scale mismatch (a system that works at small scale but breaks at large) 3. Missing safety factors (no margin for error) 4. Cost runaway (the design assumes free materials) 5. Maintenance impossibility (works on day 1, breaks on day 1000)

1Predict: which flaw category is YOUR week's work most likely to have?

I understand 5 flaw categories

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: notebook with their week's data, laptop with Gemini, the actual prototype if available. Pre-saved photos of week's prototypes are helpful.

Today's task with Gemini: feed the AI your favorite engineering project from this week, then ask it to find 5 flaws. You decide which critiques are real and which are AI hallucination - that is the lesson. Code Here - Gemini Image Gen: Web app: https://gemini.google.com (sign in with your school Google account). Click the 'Imagine' tool or paste your prompt directly with 'Generate an image of...'. Wait 5-15 seconds for the rendering. If your prompt did not produce what you wanted, refine the prompt; do not blame the model on the first try.

The Hypothesis

1Predict: how many flaws will AI find in your work? Write a number.

The Build - Full AI Critique

Pick ONE specific artifact from this week to critique. You have FOUR options - pick the one you can actually show the AI: A. PHOTO OF YOUR WIND TURBINE PROTOTYPE (Day 21) - if it's still on the bench, take a clear phone photo NOW. Side angle showing the blades + base. B. THE AI-GENERATED SUSTAINABLE BUILDING IMAGE (Day 22) - go back to your Gemini chat history from Day 22, find the image you generated, screenshot it. This works even if you don't have the prototype. C. PHOTO OF YOUR SOLAR FAN PROTOTYPE (Day 19) - if you built one, take a photo. Side angle showing the cell + motor + propeller. D. PHOTO OF A SKETCH IN YOUR WOVEN NOTEBOOK - flip to a sketch you drew this week (wind turbine blade design, solar circuit, sustainable building concept) and take a phone photo. Make sure your sketch is clearly visible. Whichever you pick, that ONE image is what you upload to Gemini in the next step. Don't describe it in words first - the AI critiques the IMAGE.

2Open Gemini at https://gemini.google.com (school account). Click the IMAGE-UPLOAD icon (paper-clip / photo icon next to the prompt box). Upload the ONE image you just picked from the list above. The image now appears at the top of your prompt.

3Below the uploaded image, type this exact prompt: 'Critique this engineering design / project. Find at least 5 specific flaws. For each, categorize as: PHYSICS (does it actually work?), SCALE (would it work at real-world size?), SAFETY (any risk to humans?), COST (realistic budget?), or MAINTENANCE (will it break in 6 months?). Be specific about WHAT in the image makes you say each flaw.' Hit send.

4Read the AI's flaws. For each, mark in your notebook: REAL FLAW, AI HALLUCINATION (the AI is wrong), or AMBIGUOUS.

5Pick the BIGGEST real flaw. Ask AI: 'How would you fix this flaw without redesigning the whole project?'

6Read the fix. Decide if it is realistic. Write your decision in your notebook.

7Build a table: Flaw, Category, Real or Hallucination, Fix Realistic?

AI sometimes 'finds' flaws that do not exist. Always verify against the data you collected. Trust your own measurements over the AI.

I got 5 AI-generated flaws

I categorized each

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

Flip the AI

1Ask AI: 'What are 3 things this design does WELL? Be specific.'

2Read the answer. Did the AI find genuine strengths, or generic praise?

3Write in your notebook: WHEN should you trust AI critique? When should you ignore it?

Real engineers calibrate their AI tool over time. They learn which prompts produce useful critique and which produce noise. This is a real skill.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

AI Critique Accuracy

Pair	Total Flaws	Real	Hallucinations	Real %
Yours	fill	fill	fill	fill
Class Avg	fill	fill	fill	fill

1What was the average percentage of REAL flaws among AI's suggestions? 50 percent? 80 percent?

In real industry, engineering teams find AI-flaw critique is about 60-70 percent useful. The remaining 30-40 percent is noise. Calibration matters.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Environmental Engineer - AI as Critic

Day in the Life of a Research Scientist

Check out this video to see a real professional in this role. QA engineers find flaws before release. The skill you practiced today is their daily work.

Quality Assurance / Reliability Engineer: What they do: find flaws BEFORE the customer does. They are the people who say 'no, you cannot ship that' and prevent expensive recalls. Entry pathway: BS in any engineering field with QA/reliability electives. Cal State LA's Industrial and Manufacturing Engineering BS is a direct route. Salary band (BLS LA 2024): entry 80,000 to 100,000. Mid-career 110,000 to 145,000. First step from where you sit today: take 'Six Sigma' yellow-belt certification (free online resources).

Reflection

1Reflection 1: What was the most painful but useful AI critique?

2Reflection 2: When have you needed someone to find your flaws before you shipped something?

3Reflection 3: Next week is the symposium. What is one thing you want to share about your work?

Week 5 Complete - climate engineers: Five days of solar physics, color filters, wind turbines, generative AI design, and AI flaw critique. You went from photons to a 1.2-million-home Mojave farm to AI-rendered sustainable buildings. Stand up. Show your favorite design. You finished Week 5. Group photo. Sign your notebook. See you next session.

I completed Week 5

My week's prototypes are critiqued

Built by Woven Learning & Technology✎ Edit

Day 24: Anatomy of a Lab Report

Research Scientist Pathway - The IMRaD Format

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

How to Structure a Research Paper (IMRaD Format)

Watch this video to set up today's thinking - this is the structure every scientific paper uses. IMRaD = Introduction, Methods, Results, and Discussion. By the end of today's lab, you will have written your own first IMRaD draft.

Foundations - IMRaD

Five weeks of data. One week to make it speak. Today you learn the structure that every scientific paper from 1950 onward has used: IMRaD.

The Hook: The Watson and Crick paper that announced the structure of DNA was 800 words long. It changed biology forever. The structure: Introduction (the question), Methods (what they did), Results (what happened), and Discussion (what it means). IMRaD. Four letters. Used in every Nature, Science, Lancet, and JAMA paper since 1953.

INTRODUCTION: the question and why it matters. (1 paragraph) METHODS: what you did. Reproducible. (1-2 paragraphs) RESULTS: the data. Tables and graphs. (1-2 paragraphs + figure) DISCUSSION: what it means. Limits and next steps. (1 paragraph) A lab report is a story. The IMRaD structure is the plot.

1Predict: which section will be hardest for you to write?

I can name the 4 IMRaD sections

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: notebook from weeks 1-5, laptop with Google Docs or Word, IMRaD template (provided by instructor), printer (optional).

Where to Write - Google Docs: Open https://docs.google.com (sign in with your school Google account). Click '+ Blank' to start a new document. Your instructor will share an IMRaD template you can copy ('File' > 'Make a copy' once they send the link). If Google Docs is blocked, use Microsoft Word: https://www.office.com (free with school account) or LibreOffice on a school laptop.

Pick Your Experiment to Write About

1Pick your single best experiment from the past 5 weeks. Write the title in your notebook. What was the question? What was the answer?

The Build - First Draft IMRaD

2Open the Lab Report Builder (above). Pick your single best experiment from the past 5 weeks. Fill in each IMRaD section: Introduction (the question + why it matters), Methods (what you did), Results (what you measured), Discussion (what it means).

3The Builder flags any section that is too thin or too vague (less than 50 words, or missing a number). Fix every flag before moving on - that is your IMRaD draft for tomorrow.

4Open the IMRaD template. Title at the top: 'Effect of [your independent variable] on [your dependent variable].'

5INTRODUCTION (3-4 sentences): What was the question? Why does it matter? What did you predict?

6METHODS (3-5 sentences): What materials? What procedure? Be specific enough that another team could reproduce.

7RESULTS (3-4 sentences + 1 table or graph): What did you measure? Show the data.

8DISCUSSION (3-4 sentences): What does it mean? What are the limits? What would you do next?

9Length: ONE page total. If yours is longer, cut. If shorter, expand.

10Trade with another pair. Read theirs. Mark sections that are clear and sections that confuse you.

Resist the urge to make the writing fancy. Simple sentences. Active voice. 'We measured X' beats 'X was measured.'

I drafted a full IMRaD report

I traded with a peer

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI as Editor

1Paste your IMRaD draft into Gemini. Prompt: 'Please review this lab report for IMRaD structure and clarity. Mark any section that is too long or unclear. Do not change the science.'

2Read the AI's feedback. Apply 1-2 of its clearest suggestions. Save your revised draft.

3Ask: 'What is the strongest sentence in my report? What is the weakest?' Read the answer.

AI is excellent at clarity and structure feedback. It is mediocre at scientific accuracy. Day 26 covers the second part.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Title and Takeaway Round

1Each pair states their title and reads ONE sentence of discussion to the class.

2Listen for diversity: how many different experiments are represented?

In real research, the title is often the most-rewritten part of a paper. Authors spend hours on those 10 words because they decide who reads it.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Research Scientist - The IMRaD Format

Day in the Life of a Research Scientist

Watch how a real research scientist organizes her day. The IMRaD work is her writing time.

Research Scientist (Industry or Academic): What they do: design experiments, run them, write papers, change fields. Entry pathway: BS minimum, MS or PhD for senior positions. Cal State LA's BS in Biology, Chemistry, or Physics is a direct entry. Apply to summer programs at UCLA, USC, JPL. Salary band (BLS LA 2024): research assistant entry 55,000 to 75,000. PhD scientist entry 95,000 to 120,000. Senior 150,000 plus. First step from where you sit today: apply to UCLA's CSULA-UCLA-USC summer research program (paid, high school open).

Reflection

1Reflection 1: Which IMRaD section was hardest?

2Reflection 2: When have you had to explain something complicated to someone simply?

3Reflection 3: What experiment from the last 5 weeks do you most want to share at the symposium?

I drafted IMRaD

I traded with a peer

Built by Woven Learning & Technology✎ Edit

Day 25: Data Visualization - Tell the Story Visually

Data Scientist Pathway - Charts That Don't Lie

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Hans Rosling — 200 Countries in 4 Minutes

Carefully view this video to ground today's hook in a real example. The classic. The data does not change. The story changes when the visualization changes.

Foundations - Three Chart Types

Yesterday you wrote about your data. Today you SHOW it. Good charts can save lives. Bad charts can lose elections. The line between the two is engineering.

The Hook: In 1858 Florence Nightingale published her 'rose chart' showing British soldiers were dying of preventable infections, not bullet wounds. The chart was so clear that Parliament passed sanitation funding within 6 months. That single chart saved tens of thousands of lives. The data had been around for years, but it took ONE good visualization to make the case.

BAR chart: comparing categories. (Tested 4 colors, which produced most voltage.) LINE chart: showing change over time or input. (Voltage at distances 10, 30, 50, 100 cm.) SCATTER plot: showing relationship between two variables. (Mass vs climb rate.) Use the wrong type and your data lies. Bar charts hide trends. Line charts hide categories. Scatter plots show correlation, NOT causation.

1Predict: which chart type best fits the experiment you wrote up yesterday?

I can name 3 chart types

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: laptop with Google Sheets, graph paper, ruler, colored pencils, notebook with their data.

Where to Build Charts - Google Sheets: Open https://sheets.google.com (school Google account). Click '+ Blank' to start a new spreadsheet. Type your data into 2 columns: column A = the thing you changed (X axis), column B = the thing you measured (Y axis). Highlight your data, then Insert > Chart. Pick the type consciously.

The Hypothesis

1Sketch what you THINK your final chart will look like. Title, axes, units, trend line. Predict the slope or pattern.

The Build - Two Charts

Cheat Sheet - Pick the Right Chart: BAR chart: comparing CATEGORIES. Use when X is a label (red filter, green filter, blue filter). LINE chart: showing CHANGE OVER TIME or input. Use when X is a number that changes (10 cm, 20 cm, 30 cm). SCATTER plot: showing RELATIONSHIP between two variables. Use when both X and Y are numbers and you want to see correlation. Decision rule: - X is a category? -> BAR. - X is a number on a scale? -> LINE. - You want to ask 'are these correlated?' -> SCATTER. Axis rule: always start Y at 0 unless the data demands otherwise. Truncated axes lie.

2Open Google Sheets. Type your data into 2 columns: independent variable (X) and dependent variable (Y).

3Highlight data. Insert > Chart. Pick the chart TYPE consciously - do not accept the default. Check that axes are labeled with units.

4Add: title, axis labels, unit labels, trendline if appropriate.

5Now hand-draw the same chart on graph paper. This forces you to understand what the chart says.

6Compare: which version is clearer? Which would convince a stranger faster?

7Write a 1-sentence CAPTION: what does this chart show?

8Bonus: redo the chart with a Y-axis that DOES NOT start at zero. Look at the same data - does the trend look bigger or smaller? This is the most common chart-lie technique.

A truncated Y-axis can make a 2 percent change look like a 50 percent change. News graphics use this trick all the time.

I built a digital chart

I built a hand-drawn chart

I wrote a caption

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Picks the Chart

1Open Gemini. Paste your data. Type: 'What chart type best displays this data? What axis ranges should I use?'

2Read the AI's suggestion. Compare to what you chose.

3Now ask: 'Could this chart be made misleading? How?' Read the answer.

Real journalists, scientists, and policy advisors all use AI for chart drafts now. The risk is AI sometimes suggests truncated axes that look more dramatic. The human always has the final call.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Chart Gallery

1Pin your hand-drawn chart on the wall. Walk through the gallery.

2Vote on three categories: most CLEAR, most BEAUTIFUL, most accidentally MISLEADING.

Category	Pair	Vote Count
Most Clear	fill	fill
Most Beautiful	fill	fill
Most Misleading	fill	fill

Real data scientists agree: clarity beats beauty. The most beautiful chart in the world is useless if a stranger cannot understand it in 5 seconds.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Data Scientist - Charts That Do Not Lie

A Day in Life as a Google Data Scientist

Watch this video to see what a real day in this career actually looks like. Google data scientists pick and build charts every day. Same skill you practiced today.

Data Scientist / Data Analyst: What they do: turn raw data into charts and insights that drive business or policy decisions. Entry pathway: BS in CS, Statistics, or Data Science. Cal State LA's BS in Computer Science with Data Science concentration is direct. Salary band (BLS LA 2024): entry 95,000 to 120,000. Mid-career 130,000 to 175,000. Senior 200,000 plus. First step from where you sit today: take Cal State LA's free Saturday Data Science Workshops, sophomore year and up.

Reflection

1Reflection 1: What surprised you about hand-drawing the chart?

2Reflection 2: When have you been tricked by a chart in real life?

3Reflection 3: For tomorrow: which chart will you bring to the symposium and why?

I built 2 versions of my chart

I caught one truncated-axis trick

Built by Woven Learning & Technology✎ Edit

Day 26: LLM Peer Review - The AI as Critic

Research Scientist Pathway - Defending Your Data

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

3 Questions to Ask Before You Believe Something (TED-Ed)

Watch this short video to see why today's topic matters before we dive in. The 3 questions in this TED-Ed are exactly the questions a real peer reviewer (or an AI peer reviewer) will ask of your draft. Today you will turn those questions on your own work.

Foundations - The Three Critic Roles

Yesterday you made the data look good. Today the AI tries to tear it apart. Real science survives review by skeptics. So will yours.

The Hook: When you submit a paper to Nature, three reviewers tear it apart. They look for: bad statistics, missing controls, unsupported claims, logical gaps. Most papers get REJECTED. The few that survive get rewritten 4 to 8 times. This is rigorous. It is also why science can be trusted.

STATISTICAL critic: are the numbers right? LOGICAL critic: do the conclusions follow from the data? METHODS critic: was the experiment well-designed? A single AI can play all three roles. You will use it for each.

1Predict: which kind of critique will hit your work hardest?

I understand the 3 critic roles

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: laptop with Gemini or Claude, IMRaD draft from Day 24 (revised based on Day 25 charts), notebook.

Today's task with the LLM (Gemini or Claude): paste your IMRaD draft and run a structured 4-pass peer review (clarity, logic, evidence, hostile-audience). The AI plays peer reviewer; you decide what feedback is real. Code Here - LLM for Peer Review: Web app: https://gemini.google.com or https://claude.ai (sign in with your school account). Start a NEW chat for each critic role (statistical, logical, methods) so the AI does not contaminate one critique with another. Paste your full IMRaD draft each time. Keep the AI's response open in a tab so you can quote it in your notebook.

The Hypothesis

1Predict: how many of AI's critiques will be FAIR? How many will be unfair or wrong?

The Build - Three Critiques

2Critique 1 (Statistical). Paste your draft. Prompt: 'Review this lab report as a statistical critic. Are the numbers, percent calculations, and trend claims sound?' Take notes.

3Critique 2 (Logical). New chat. Paste draft. Prompt: 'Review as a logical critic. Do the conclusions in the Discussion follow from the Results? Find at least 2 gaps.'

4Critique 3 (Methods). New chat. Paste draft. Prompt: 'Review as an experimental design critic. Was the experiment controlled? What confounding variables were missed?'

5For each critique, list every point the AI made. Mark each as VALID, INVALID, or AMBIGUOUS.

6Pick the 2 most VALID critiques. Revise your IMRaD draft to address them. Save as version 2.

7Build a table: Critique Type, AI Point, Verdict, Revised?

Some AI critiques will be wrong. Do NOT change your work to satisfy a wrong critic. Defend with data.

I ran 3 AI critiques

I revised based on the strong ones

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

Why this app: in Phase 2 you used the Lab Report Builder to get AI peer review on your IMRaD draft. Now reuse the SAME app with a new lens - feed in your near-final draft and switch the prompt from 'fix my writing' to 'predict the hardest questions a senior researcher will throw at me at the symposium tomorrow.' Same tool, sharper purpose. Open it, paste your draft, and run the steps below.

1Open the Builder in Peer Review mode (above). Paste your IMRaD draft. The app sends it to the LLM with structured prompts: 'find clarity issues', 'find logic gaps', 'find evidence gaps', 'predict the 5 hardest questions a senior researcher would ask'.

2Read the AI's 5 predicted questions. For each: write a 1-2 sentence answer in your notebook. Practice saying them out loud - tomorrow at the symposium, ONE of these questions will probably get asked.

AI Predicts Symposium Questions

3Same chat as Critique 2 (logical). Type: 'Based on this lab report, what 5 hardest questions might a senior researcher ask the student presenter?'

4Read the 5 questions. Write them in your notebook.

5Draft a 1-2 sentence answer to each. Practice saying them out loud.

Real researchers prepare for hostile audiences. The hardest question is the one you have not thought of - which is exactly what AI just gave you.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class Critique Validity

Critic Type	Total Points	Valid	Invalid	Valid %
Statistical	fill	fill	fill	fill
Logical	fill	fill	fill	fill
Methods	fill	fill	fill	fill

1Which critic role gave the most useful feedback? Why?

Real journals rate reviewers on this same metric. A reviewer who finds 80 percent valid issues is gold; one who finds 30 percent gets dropped from the rotation.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Research Scientist - Defending Your Data

What is a Data Scientist? A Day in the Life

Check out this video to see a real professional in this role. Defending data through peer review IS what working scientists do.

Research Scientist with AI Workflow: What they do: lead lab teams, write grants, defend results in peer review. Use AI to accelerate but always own the result. Entry pathway: PhD typical for senior. BS or MS for research associate roles. Cal State LA + UCLA/USC PhD program is a strong LA route. Salary band (BLS LA 2024): research associate entry 70,000 to 90,000. PhD scientist entry 95,000 to 120,000. PI / lab head 175,000 plus. First step from where you sit today: ask Cal State LA's BIO/CHEM/PHYS departments about high school summer lab assistant positions.

Reflection

1Reflection 1: Which AI critique stung the most? Was it valid?

2Reflection 2: When in life have you needed to defend an idea?

3Reflection 3: What is your strongest single result from these 5 weeks?

My IMRaD is peer-reviewed

I have 5 anticipated questions ready

Built by Woven Learning & Technology✎ Edit

Day 27: Symposium Dry Run

Research Scientist Pathway - The 5-Minute Defense

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

Yesterday you ran AI peer review on your lab report. Today you practice the 5-minute presentation you'll give to Cal State LA tomorrow.

How to Sound Smart in a TEDx Talk (Will Stephen)

Watch the video to set up today's thinking - this is the real-world story behind the lab. A funny but real talk on how to deliver a presentation. Tomorrow you stand up and present your data.

Foundations - The 5-Minute Structure

Tomorrow you present to real Cal State LA faculty and your families. Today is dress rehearsal. Five minutes per team. Question and answer after.

The Hook: A TED talk is 18 minutes max. A scientific conference talk is 12 minutes. A research showcase answer is 30 to 90 seconds. The rule: the BIGGER the audience, the SHORTER the talk. Tomorrow you have 5 minutes per pair. Make every word count.

Minute 1: HOOK + question. (Why does this matter?) Minute 2: Methods. (What did you do?) Minute 3: Results. (Show the data + chart.) Minute 4: Discussion. (What does it mean? Limits?) Minute 5: Pathway connection + thank you. Leave 2 minutes for Q+A after.

1Predict: which minute will feel the longest when you are presenting?

I can name the 5-minute structure

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: IMRaD report (final v3), main chart printed or on screen, notebook with anticipated questions, stopwatch, partner pair.

What You Need Open: Your IMRaD draft from Day 24 + 26 (Google Docs). Your main chart from Day 25 (printed OR on screen). A stopwatch on your phone. A partner pair to present to. No new software. Today is voice + chart + 5 minutes.

The Hypothesis

1Predict your dry-run time. Will you go OVER or UNDER 5 minutes?

The Build - Two Dry Runs

Cheat Sheet - The 5-Minute Talk: Time budget (rehearse with a stopwatch): Minute 1: HOOK. One real-world story (the same one in your IMRaD intro). End with your research question. Minute 2: METHODS. What you actually did. Show the tape arena, the drone, the cuff. Be specific. Minute 3: RESULTS. Show the data. Pause on your chart for 5 seconds. Read the headline number out loud. Minute 4: DISCUSSION. What it means. One limitation. One next step. Minute 5: CAREER + thanks. The pathway it connects to. Thank the audience. Two speaking rules: - 'Um' fillers: replace with a SILENT pause. Pauses make you sound confident. - Eye contact: pick 3 friendly faces in the room and rotate. Do not stare at a slide.

2Dry Run 1: Pair A presents to Pair B for 5 minutes. Pair B times it and asks 2 questions after. Pair A answers.

3Pair B gives feedback: what was clear, what was confusing, what was missing.

4Swap: Pair B presents to Pair A. Same protocol.

5BREAK: 5 minutes. Both pairs revise their talk based on feedback.

6Dry Run 2: Same partners. Both pairs present again. Time it. Improvement?

7Note in your notebook: 1 thing you fixed, 1 thing still rough.

Going OVER 5 minutes is the most common mistake. Cut your hook by half if needed. The data is the star.

I did 2 dry runs

I got feedback

I revised between them

3. Build (Vibe-Code Your Companion App)

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

You've Used 15 Apps This Summer - Today You Build Your Own

In 6 weeks, your team has used FIFTEEN different interactive apps in this workshop: Robotics: Sphero Pathfinder (Day 3), Sphero Pathfinder Showdown (Day 4) Drones: Drone Lift Calculator (Days 6, 7, 9), Drone Battery Predictor (Day 8), ArUco Lux Simulator (Day 13) Medicine: Vital Signs Interpreter (Days 14, 15), Diagnostic Pattern Matcher (Day 18) Energy: Solar Filter Plotter (Day 20), Wind Turbine Grapher (Day 21) Writing: Lab Report Builder (Days 24, 26) Every single one of these was vibe-coded by a developer with the same tools you have today: Gemini Canvas + a deploy host like Netlify. Today YOU build one. Tomorrow at the symposium, your URL goes on the whiteboard, the audience pulls it up on their phones, and your app becomes part of your presentation.

Why we do this: a static IMRaD report tells the audience your data. A LIVE INTERACTIVE APP lets them experience it. Real research labs publish interactive supplements alongside papers (papers + GitHub repo + live demo) - the demo is what gets shared on Twitter / LinkedIn / family group chats. Today you make your own.

Tutorial - Vibe Coding with Gemini Canvas

Vibe Coding with Gemini Canvas (Google for Education)

Carefully watch this short tutorial first if you have not built with Canvas before. Same flow as Day 22's Gemini Canvas exercise - this time you ship the result instead of just designing.

8 Apps in 15 Minutes - Gemini Canvas (Teacher's Tech)

Skim this longer tutorial for IDEAS - it shows 8 different mini-apps built with Canvas in under 15 minutes. Pause and copy any pattern that feels close to what you want.

Pick Your App Idea (5 min hard limit - choose ONE)

Pick ONE category. The app you build TODAY will go up on the whiteboard tomorrow alongside your IMRaD talk - so it MUST connect to your symposium presentation. Don't try to be impressive. Try to be USEFUL. What ONE specific thing from your symposium do you want the audience to UNDERSTAND in 30 seconds? A. INTERACTIVE CHART of your data: take your Day 25 chart and let viewers hover, click, filter. Example: drone payload vs flight time, slider for payload, real-time prediction line. B. SIMULATOR of your finding: a tiny version of your experiment users can play with. Example: solar cell output simulator (slide light intensity, see voltage match your real-world curve), or a vital signs interpreter for the patient case in your talk. C. QUIZ that teaches your topic: 5 multiple-choice questions related to YOUR symposium project. Example: 'Which drone payload was MY lift threshold?' with hints from YOUR data. Audience tests their understanding of your talk. D. CALCULATOR for your finding: a tool the audience can use AFTER your talk. Example: 'How many of MY wind turbines does your house need?' or 'Plug in YOUR house orientation to see solar-panel output' - directly from your experiment's math. The app must show YOUR data, simulate YOUR experiment, quiz YOUR topic, or use YOUR formulas. If you cannot connect it to your symposium presentation in one sentence ('this app helps the audience [verb] my [topic] by [doing something]'), pick a different category. The whole point: tomorrow your URL goes up on the whiteboard, the audience pulls it up on their phones, and your app becomes a living extension of your 5-minute talk.

1Write your ONE app idea in your notebook in this exact format: 'My app will let the audience [verb] [my-symposium-topic-noun] so they can [outcome that connects to my talk].' Example: 'My app will let the audience adjust drone payload so they can predict flight time using MY Day 8 data.' If your sentence doesn't mention your symposium topic + your data, rewrite it.

Build Your App (30 min hard limit)

2Open Gemini Canvas (the same Gemini you used Day 22-23). If you have a Day 22 chat history with Gemini Canvas open, you can reuse it. Otherwise: go to https://gemini.google.com, click 'Canvas' in the prompt area, hit New.

3Write your build prompt. Template: 'Build a single-page HTML app with embedded JavaScript and CSS. CONTEXT: I just finished a 6-week research project on [your symposium topic]. Tomorrow I present my findings and want a live companion app the audience can use during my talk. The app should: [your one-sentence idea from above]. The user should be able to: [the 1-2 main interactions]. Use my actual data values: [paste 4-6 numbers from your Day 25 chart / table here]. Use only inline CSS + JS - no external CDNs that might be blocked at school. Use clean modern colors (teal + navy + gold accents to match my workshop branding). When the user opens it, the app shows: a friendly title naming my topic + 1-line description of my experiment + the interactive piece below.' Hit send. The kid who pastes their REAL data values gets a way better app than the kid who keeps it generic.

4Read the generated code. Click 'Run' / 'Preview' to see it work. Test the interactions. Does it do what you said? If not: in the same chat, write 'Iterate: [specific change you want]' and send. Repeat up to 3 times.

5Stop iterating at 30 minutes regardless of state. Better to ship a simple working app than chase perfection. Click the DOWNLOAD button (top-right of the Canvas pane) to save your file as something like 'app.html' to your Downloads folder.

Deploy to Netlify (5 min)

Netlify Drag-and-Drop Deploy Tutorial

Watch this short Netlify-official video first if you've never deployed an app before. The whole flow is exactly: drag .html file onto netlify.com/drop, get a URL.

6Open https://app.netlify.com/drop in a new tab. Sign in using the SHARED CLASSROOM Netlify credentials your instructor projected on the whiteboard.

7Drag your downloaded .html file from your Downloads folder onto the netlify.com/drop area. Wait 10-30 seconds. You will see 'Site is live!' and a random URL like 'https://elegant-kelpie-12345.netlify.app/'.

8Click the URL to open your app in a fresh tab. Test it. Does it work the same as it did in Gemini Canvas? If not: re-download from Gemini and re-drag. The new version replaces the old at the same URL.

9Open the URL on your PHONE (not your laptop). Confirm it loads on mobile. The audience tomorrow will use phones - if it doesn't render, fix it now (ask Gemini: 'Make this responsive on mobile').

Share Your URL With the Class

10Walk to the whiteboard. Find the URL TABLE your instructor pre-drew. Add YOUR row: Pair Name | App Name | What it does (1 line) | URL. Write your URL clearly - the audience tomorrow will type it in.

11Optional bonus: if you want to remember the URL, you can rename the random subdomain in the Netlify dashboard ('Site settings' > 'Change site name' > pick a memorable name). Skip if short on time.

12Plan how you'll use the app TOMORROW: will you mention the URL during your 5-min talk? Show it on a slide? Tell the audience to pull it up on their phones at the end? Write your plan in your notebook.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Class App Gallery - Test Each Other's Apps

Copy this table into your Woven notebook BEFORE the gallery walk. Sketch the column headers neatly on a fresh page. As you visit each app, fill in your notebook copy. Your instructor's whiteboard URL table is the master list.

Class App Gallery

Pair	App Name	URL	Did it work?	One thing you liked	One suggestion
Yours
App 2
App 3
App 4
App 5

1Pull up the WHITEBOARD URL TABLE on your laptop or phone (the 14-15 URLs your classmates wrote). Pick 4 OTHER pairs' apps to visit (not your own).

2Visit each URL. Spend 2 minutes per app. Try every interaction. In your notebook gallery table, record: did it work? one thing you liked, one suggestion.

3Find the pair whose app you found most useful. Tell them in person why - 'I liked your app because [specific].' This is the kind of feedback you want at the symposium tomorrow.

Tomorrow at the symposium, your URL goes on the whiteboard alongside everyone else's. The audience pulls them up on their phones during or after talks. The 5-minute symposium presentation is just the beginning - your live URL is the take-home. Real research labs ship papers AND code AND demos. So did you. So did everyone in this room.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Research Scientist - The 5-Minute Defense

How to Sound Smart in a TEDx Talk (Career Connection)

Watch this video to see what a real day in this career actually looks like. Already on Day 27 - a real-world reminder of presentation craft.

Research Scientist (Public Communication): What they do: present results to non-scientists - donors, policymakers, the public. The most senior scientists are also great communicators. Entry pathway: any STEM BS plus practice. Cal State LA's STEM Ambassador program trains undergrads in public science communication. Salary band: communication-focused PhDs in STEM make 110,000 to 180,000 in LA. Pure-comm professionals (science writers) 75,000 to 120,000. First step from where you sit today: enter the Cal State LA Three-Minute Thesis competition (open to undergrads, watchable on YouTube).

Reflection

1Reflection 1: What is the strongest single sentence in your talk?

2Reflection 2: When have you had to perform under pressure before?

3Reflection 3: What is the ONE thing you want the audience to remember tomorrow?

I rehearsed twice

I am ready for tomorrow

Built by Woven Learning & Technology✎ Edit

Day 28: The Symposium - Present to Cal State LA

Research Scientist Pathway - The Public Defense

1. Spark

Woven notebook: open your notebook now. Use it to capture every prediction, partner discussion, and question that comes up in this phase. Your notebook is the record of your thinking.

All week you have practiced. Today you present.

How to Sound Smart in a TEDx Talk (Will Stephen)

Watch this once more before you go on. Same TEDx talk from Day 27 - this time as your last reminder right before the symposium. Then go.

Final Foundations - Symposium Etiquette

Six weeks. 27 days of labs, drones, vitals, solar cells, AI critiques. Today you present to Cal State LA faculty and your families. This is the moment your work meets its audience.

The Hook: Every scientist on Earth had a day like today: their first public showcase of their own research. Albert Einstein gave his at age 26 to the patent office in Bern. Marie Curie at 30 to the French Academy. Jennifer Doudna at 24 to her PhD committee. It feels huge. It is huge. It is also just a 5-minute talk you have already rehearsed twice.

BEFORE you present: shake hands with the moderator. Smile. DURING: stand still. Make eye contact with 3 people. Pause on your chart for 5 seconds. AFTER: thank the audience. If you cannot answer a question, say 'I do not know' and offer to follow up. That last sentence is what experienced researchers say. Faking certainty kills credibility.

1Predict: which minute of the symposium will feel longest?

I am ready

2. Lab

Woven notebook: keep your notebook open as you build. Record your hypothesis, every measurement and observation, and what surprised you. Your notebook is your lab record for today.

Materials per pair: IMRaD final draft (printed AND on a USB drive as backup), main chart printed on 8.5 x 11 paper, water bottle, notebook for taking notes during others' talks. Audience: Cal State LA faculty, families, program staff. Order announced 5 minutes before start.

The Live Symposium

1AV check: load your slide or handout. Test the projector. 5 minutes before symposium starts.

2Audience seated. Welcome from program director. First pair invited up.

3WHEN IT IS YOUR TURN: walk to the front. Take a breath. Smile. Begin with your hook sentence.

4Present your IMRaD in 5 minutes. Pause on the chart. Make eye contact.

5Mention your COMPANION APP URL once during your talk - either in the methods slide ('I built an interactive demo of this experiment - it is live at [URL]') or as your closing line ('Pull up [URL] on your phones if you want to see the data move'). The whiteboard has all 15 URLs projected so the audience can browse during transitions.

6Q+A: 2 minutes. Listen fully before answering. 'I do not know - I would test that next' is a real answer.

7After your slot: sit down. Listen to other teams. Take notes.

8After all pairs: closing remarks from program director. Group photo. Certificates.

If you blank in the middle, just say 'one moment' and look at your notes. The audience will wait. Real scientists do this all the time.

I presented my work

I answered Q+A

I listened to my classmates

3. AI Check

Woven notebook: as the AI helps or fails, write down exactly what you fed it and what it gave back. The trail of prompts and outputs IS your data.

AI Across the 6 Weeks

1Across all 28 days, AI was used for: code generation, model fitting, image generation, peer review, hostile-audience prep. Write down which AI use was MOST useful for you.

2Write down ONE AI use that was MISLEADING - where the AI gave a confident wrong answer.

3In one sentence: when should a researcher trust AI, and when should they not?

Your answer to that one sentence is your AI literacy. Most adults working today cannot answer it. You can.

4. Class Data

Woven notebook: pull your data into a clean table. Write the trend you see in one sentence. If you cannot describe it in one sentence, you do not understand it yet.

Program Wrap

Metric	Count
Pairs presented	fill
Audience questions answered	fill
Different research topics	fill
AI tools used across the program	6+

1Find one teammate or peer whose work surprised you. Tell them in person.

Most LA high schoolers will never present research to a college audience before they are 19. You did it at 14 to 17. Carry that.

5. Wrap

Woven notebook: answer the reflection prompts in writing before you leave. Tomorrow's session starts where today's notebook ends.

Pathway: Research Scientist - The Public Defense

A Week in the Life of a Research Scientist

Watch this video to see what a real day in this career actually looks like. Imagine yourself here. The weeks you saw across this summer scale up to research lab life.

You, in 5 Years: Cal State LA Upward Bound students who present research at this symposium are 2.5x more likely to enroll in a 4-year college (federal UB longitudinal data, 2024). The pathways you saw this summer are real: - Robotics Engineer: 90,000 to 200,000+ - Aeronautical Engineer: 84,000 to 175,000+ - ML / CV Engineer: 110,000 to 250,000+ - EMT to RN to MD: 50,000 to 300,000+ - Solar/Wind Engineer: 90,000 to 175,000+ - Research Scientist: 70,000 to 175,000+ First step from today: keep your notebook. Apply to Cal State LA's Early Entry Program in fall. Continue Upward Bound through your senior year. The summer you just finished is part of the case.

Reflection

1Reflection 1: What is the ONE moment from these 6 weeks you will remember in 10 years?

2Reflection 2: Which pathway from the program excites you most? Why?

3Reflection 3: What is the first thing you will do tomorrow with what you learned this summer?

Hold on to your notebook. It is the most concrete proof of your work. College admissions officers will ask about it. Keep it for at least 4 years.

Symposium Eve - You're Ready: You've practiced your 5-minute presentation twice. You've heard the AI's hardest questions. Tomorrow Cal State LA staff and your families fill the room. Stand up. Say your single strongest sentence out loud. Sleep well. You're ready. Group photo. Sign your notebook. See you next session.

I completed the symposium

I completed the program

I will keep my notebook

Built by Woven Learning & Technology✎ Edit