llm-architecture

Agentic RAG: Adding Reasoning to Retrieval

Moving beyond simple vector search. How Agentic RAG uses multi-step reasoning, query decomposition, and corrective feedback to answer complex questions.