# Jeffery Nketiah

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Jeffery Nketiah](https://cerebralvalley.ai/u/jefferynketiah20)
- **GitHub:** https://github.com/jefferynketiah-gif/LaboraVR-MetaAI
- **Demo video:** https://youtu.be/u-1wBPbxWDc
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2595

LaboraVR is a virtual-reality chemistry lab on Meta Quest for Cambridge IGCSE students. Every year thousands of them sit Paper 6, the "Alternative to Practical" exam, describing chemistry they have never seen, because their school has no lab, no chemicals or no time. In LaboraVR they do the tests with their own hands, and Meta's Muse models make it a lab that a teacher can direct and that examines the student the way the real paper does.

Three AI features, all running on the Quest:

AI Practical Architect. The teacher (or student) says what they want, for example "a practical to tell aluminium from zinc". Muse Spark designs the practical: the tubes, reagents and steps. A deterministic checker then tests the design against the teacher-written specification of every IGCSE test (no chemistry is in the code). If the design is wrong (a reagent that cannot tell two ions apart, a missing acidification step) the checker rejects it with the reason, and Muse corrects itself. The accepted practical is built on the bench in front of the student, and saved, so the next class runs it offline with no AI cost.
Spoken observation examiner. After each test the student says what they saw. Muse Voice Transcribe writes it down, Muse Wonder extracts each claim, and the marking is done by rules against the real mark-scheme wording, like a Paper 6 examiner: "white precipitate" scores, "it went cloudy" does not, and the student is told why.
Planning examiner. The student says their plan for an investigation aloud; Muse maps it onto the plan's mark points and the same rules mark it.
Mixed reality puts the practical on the student's real desk, and the teacher gets a class report of what each student said and scored. When a practical is finished, one tap has Muse design the next practical from that student's own mistakes, checked against the specification like every other design. The problem it solves: students lose marks on Paper 6 because they have never seen the reactions and have never had to put observations into examiner's words. LaboraVR gives them both, without a lab.

Built during the hackathon (26 October to 4 November 2026): everything above that uses AI. The Practical Architect and its checker, the spoken observation and planning examiners, the next practical from the student's mistakes, speak first (no answer choices until the student has said what they saw), mixed reality on the real desk, and the class report. The VR lab itself (the room, the apparatus and the guided lessons) existed before the event.

LaboraVR is not affiliated with or endorsed by Cambridge University Press & Assessment. "Cambridge IGCSE" is their trademark, used here only to say which exam the app prepares students for.

## More from Meta Global AI Developer Hackathon

- [StormGate-Zero](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2592)
- [Infinity](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2593)
- [Pawprint Club](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2594)
- [Fight Club](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2596)
- [SignBridge](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2597)
- [Ode: Give Muse a body](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2598)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2595. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt