3 projects
geulpu
A deterministic local LLM engine and debugger — bit-reproducible GGUF inference on your own GPU, written in the geul language (Windows x64 and Linux x86-64 apps, CUDA or Vulkan)
geul
글 — 한국어 문법(SOV·조사)으로 쓰는 자체호스팅 시스템 언어의 컴파일러 (Windows x64)
entail-ai
Keep what a value means intact across LLM inference-stack boundaries: declare it, check it against the data, and resolve a mismatch before it becomes a silent wrong answer