# Bignesh Das

- **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:** [BIGNESH DAS](https://cerebralvalley.ai/u/Bignesh)
- **GitHub:** https://github.com/Bignesh14/glimmer-m5-selftune
- **Demo video:** https://www.youtube.com/watch?v=h_H0wiRkM7U
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1745

Can an AI model make its own inference faster, safely? Muse Glimmer 30B (Meta, 4-bit, MLX) runs on a Mac Studio M5 Max. I built a referee: a safety gate, a bit-exact correctness check against a golden reference, and replicated, interleaved A/B speed races with rules fixed before any result. With it, Glimmer now generates 5–21% faster on its own 19 recorded self-tuning prompts (median +12.8%), 17% faster on code, and faster on every task of a Spec-Bench subset (1.15x overall). That's 252/252 paired runs faster, with no extra memory and identical text in fp32 arithmetic. Two pieces: H8, prompt-lookup decoding (Saxena 2023) made exact for Glimmer's sliding-window cache, which mlx-lm's own speculative decoder refuses to roll back; and H9, a fix for an MLX attention-kernel limit (query length x GQA factor <= 32) that the referee found when gains collapsed on long contexts. When Glimmer was asked to fix that itself, it found the speed-up 10 times in 16 attempts, but every fast version let tokens peek at the answers being checked. A speed-only reward would have accepted all 10; the referee rejected every one. Try it with one command on any Apple Silicon Mac (./run_demo.sh --small for 8 GB), or ./run_demo.sh --ui for a web page where confirmed guesses show in green. The README states every limit (smaller gains on prose and small Macs).

## More from Meta Global AI Developer Hackathon

- [Imageek](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1742)
- [cravour.com](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1743)
- [OrderThread](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1744)
- [CHAMP SUPER ANTELLIGENT](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1746)
- [Prabash Vijayanga](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1747)
- [Obfinity](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1748)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1745. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt