Pinned Loading

1. hydra hydra Public

   A stealth-automation primitive on Camoufox: auto-discovers a site's internal API and self-heals when detection breaks — usable by anyone automating the web.

   Python 1

2. flint flint Public

   A from-scratch int4 decode megakernel for Granite-4.1 — and the measured finding that batch-1 on H100 is latency-bound, not bandwidth-bound.

   Python

3. leankv leankv Public

   Stacking activation sparsity and KV cache quantization for LLM inference throughput(B=1) and larger context window.

   Python

4. LLM-Doc LLM-Doc Public

   Layer-wise debugger for quantized LLMs: is it broken, where, why, and how to fix it.

   Python 2 1

5. Prospector Prospector Public

   Automated job prospecting pipeline — stealth scraping, LLM-powered fit scoring & outreach drafting, decision-maker discovery, Discord delivery. Human-in-the-loop.

   Python

6. aeonmindai/runcrate-py aeonmindai/runcrate-py Public

   Official Python SDK for Runcrate — manage GPU instances and run model inference

   Python 1