Leeroopedia

关于

Leeroopedia is an expert-level knowledge base for the entire ML & AI stack, designed to empower AI coding agents. It covers over 1000 frameworks and libraries, ranging from fine-tuning and distributed training to inference serving, GPU kernel optimization, agent building, and RAG pipelines. As an MCP server, it seamlessly integrates with various AI coding agents like Claude Code, Cursor, and OpenAI Codex, transforming them into proficient ML/AI expert engineers capable of tackling complex development tasks with best practices.

主要功能

• Comprehensive ML & AI knowledge base spanning 1000+ frameworks and libraries
• Seamless integration with popular AI coding agents (e.g., Claude Code, Cursor, OpenAI Codex)
• Significant performance improvements for agents on ML tasks, validated by benchmarks
• Optional skill file to optimize agent tool selection and workflow
• Supports both hosted server and local installation via uvx
• 5 GitHub stars

使用案例

• Optimizing ML inference and writing custom CUDA/Triton kernels
• Performing end-to-end LLM post-training workflows (SFT, DPO, LoRA merge)
• Building self-evolving RAG services that automatically improve over time