Lightweight, auditable Python code agent (~1500 LOC) — ReAct + Planner + Reflexion + Hybrid RAG, with SWE-bench Lite eval and trace replay.

Branch-based learning repo for Agentic AI and LLM engineering with LangGraph: ReAct tool calling, Agentic RAG, MCP, Reflection, Reflexion, Human-in-the-loop memory, and DeepAgents examples.

AI Agent Architectures is the complex structures that shape how machines perceive, reason, and act in their environments in the pursuit of autonomous intelligence.