# amk

- **Event:** [AIE Code Summit 2026 Hackathon](https://cerebralvalley.ai/e/aie-code-2026-hackathon)
- **When:** Sat, Nov 7 at 9:00 AM – 10:00 PM (PST)
- **Where:** San Francisco, CA
- **Team:** [Subrata Goswami](https://cerebralvalley.ai/u/whatdhack)
- **GitHub:** https://github.com/whatdhack/amusingkernels
- **Demo video:** https://github.com/whatdhack/amusingkernels/blob/main/assets/demo.gif
- **Gallery:** https://cerebralvalley.ai/e/aie-code-2026-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/aie-code-2026-hackathon/hackathon/gallery/16

amusingkernels is an LLM-driven agent that writes, debugs, and optimizes GPU kernels for NVIDIA's SOL-ExecBench benchmark, producing leaderboard-ready
  submissions. Muse drives a tool-using loop — read the problem, write a kernel, validate its signature, evaluate on a local GPU or a Modal B200, read
  back the trace — while Gemini acts as second-opinion critic on failures and fresh-eyes proposer when the search stalls, and KernelWiki contributes
  retrieved optimization techniques matched to the run's measured symptoms. At its heart is an exercise in software archaeology: a GPU-less analytical
  model of the NVIDIA B200, reconstructed without access to the vendor's proprietary timing models — a timed Petri net of the TMA → tcgen05.mma →
  epilogue pipeline assembled from datasheet peaks and CUTLASS pipeline semantics, bit-exact functional models of narrow-precision tensor arithmetic,
  and hard architectural gates for shared memory, TMEM, and instruction shapes, all pieced together from public shards. Every Modal B200 evaluation is
  logged as a predicted-vs-measured pair, so each real trace carbon-dates the reconstruction: agreement is reported as latency ratios, rank correlation,
  and bottleneck confusion, and the residuals drive calibration of the model's uncalibrated constants. Early results validate the approach end to end:
  seven frozen L1 solutions, a 10-turn dual-model FlashInfer-Bench GEMM campaign, and a first calibrated comparison (123µs predicted vs 50.5µs measured)
  closing the loop between excavation and evidence.

## More from AIE Code Summit 2026 Hackathon

- [STRATA.](https://cerebralvalley.ai/e/aie-code-2026-hackathon/hackathon/gallery/13)
- [Trust Issues](https://cerebralvalley.ai/e/aie-code-2026-hackathon/hackathon/gallery/14)
- [Never Twice](https://cerebralvalley.ai/e/aie-code-2026-hackathon/hackathon/gallery/15)
- [Theseus](https://cerebralvalley.ai/e/aie-code-2026-hackathon/hackathon/gallery/17)
- [Dojo](https://cerebralvalley.ai/e/aie-code-2026-hackathon/hackathon/gallery/18)
- [Thought to Reality - Dylan Ler](https://cerebralvalley.ai/e/aie-code-2026-hackathon/hackathon/gallery/19)

---

Markdown version of https://cerebralvalley.ai/e/aie-code-2026-hackathon/hackathon/gallery/16. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt