I do research on AI
A straightforward method for training your LLM, from downloading data to generating text.
35 production-grade agentic AI architectures (Reflexion, LATS, GraphRAG, MemGPT, Voyager, BrowserAgent, ...) — a Python library and runnable textbook with multi-provider LLM support and a 17-task benchmark leaderboard.
A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.
Implementation of all RL algorithms in a simpler way
Core 7 layers of production grade agentic system
Building DeepSeek R1 from Scratch
Handling 10M+ docs using RAG with zero hallucinatons
