Llm App Patterns

If you're building LLM applications beyond simple prompts, this gives you the architectural patterns you actually need. It covers RAG pipeline design with practical chunking strategies and retrieval methods (semantic, hybrid, multi-query), agent architectures like ReAct and function calling, and the operational patterns for running these systems in production. The code examples are production-oriented, not toy demos. Useful when you're past the prototype phase and need to make decisions about vector databases, embedding models, or which agent pattern fits your use case. It's basically the architecture documentation you'd build internally after shipping a few LLM features, packaged as reusable patterns.