# Weerakorn_TH

- **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:** [Weerakorn Leartsirisorn](https://cerebralvalley.ai/u/weerakorn_TH)
- **GitHub:** https://github.com/Weerakorn-Leartsirisorn/lecture-lens
- **Demo video:** https://youtu.be/F1pZ_OWEB6I
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4415

Lecture Lens turns lecture slides into a study kit students can trust: a bilingual Thai–English summary, a practice quiz, and flashcards, where every answer links back to the exact slide it came from.

The problem: at Chulalongkorn University, lectures mix Thai and English and slide decks run 60+ pages. AI summarizers save time but invent facts, and students can't tell which generated questions are actually covered in their course material.

How it works:
- Upload a lecture PDF. It is split into chunks and Muse Spark 1.3 (Meta Model API, Files API + Responses API) reads every page in parallel, including tables, formulas and diagrams.
- A structured summary in Thai and English follows the lecture's own order, with a key-term glossary. Technical terms stay in English, the way students use them in class.
- Every quiz question and flashcard must cite a source page and quote the slide verbatim. The app then verifies that the quote really appears on that page, and shows verified items first (15/15 flashcards verified on a 68-page Decision Tree lecture).
- Click any page number to open the original slide beside the answer.

Built with FastAPI, Pydantic-validated JSON outputs with automatic retries, and a vanilla JS frontend polished with Muse Code.

## More from Meta Global AI Developer Hackathon

- [Donald O. Ebuga](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4412)
- [Introvert Devs](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4413)
- [Sentinel Force](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4414)
- [Tellhouse](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4416)
- [Newport1 team](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4417)
- [Codeminister](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4418)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4415. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt