llm_interview_note collects concepts, knowledge, and interview questions for large language model algorithm and application engineers. The material is compiled from public resources and is intended for study and interview preparation. The author also maintains tiny-llm-zh for hands-on learning under limited hardware resources.

The README highlights tiny-llm-zh for building a small Chinese LLM from scratch and learning pretraining, fine-tuning, and reinforcement learning; tiny-rag for building a simple RAG system with multi-path retrieval and reranking; tiny-mcp for implementing MCP servers and clients with prompts and function calling; and llama3-from-scratch-zh for implementing Llama 3 from scratch while supporting Meta weights and local debugging on a 16 GB machine.