Session by session
- Session 31 · Ask & understandWeek 16
What model scores do and do not mean
Evidence: a question, sketch or new vocabulary
- Session 32 · Build & designWeek 16
Record predicted labels and confidence scores
Evidence: a design draft, dataset or build
- Session 33 · ImplementWeek 17
Compare two acceptance thresholds
Evidence: a working version, explained once
- Session 34 · Test & improveWeek 17
Count false accepts and false rejects
Evidence: a test log with at least one failure
- Session 35 · Explain & reflectWeek 18
Choose an abstain rule and justify it
Evidence: an individual explanation
- P03 model
- held-out images
- local table
Twenty test cases; never label confidence as guaranteed probability.
A model, an evaluation study or an AI-checking workflow.
Completion needs the artefact, an honest test log, an individual explanation and no open safety or privacy issue.
Projects in the same block
- RoboticsG07-P05
Distance Display
A display that reports valid distance or an explicit invalid reading.
Sessions 21-25Builder and up - RoboticsG07-P06
Smart Night Lamp
A motion-and-light lamp model with a manual override.
Sessions 26-30Builder and up - AIG07-P07
Confidence & Error Lab
A threshold comparison table with a visible no-decision outcome.
Sessions 31-35Builder and up - IntegratedG07-P08
AI Waste-Gate Adviser
A supervised waste-sorting gate with an explicit unknown route.
Sessions 36-40Builder and up
For school leadersChoose a starting point.
Build from evidence.
Pick the classes and a plan. We map the timetable, kit and safety checks with you, then pilot one class first.
- Prospectus and class-wise plan
- Kit and readiness check
- Pilot one class first
- Evidence at every milestone
Let's plan your pilot.
Share a few details and we will send the right plan for your classes.