Kemal Kondakçı
← ProjectsMulti-Agent AI System

Multi-Agent AI System

A full-stack system where multiple AI agents collaborate to solve a task. State-aware agent flows orchestrated with LangGraph; Python (FastAPI) backend + React frontend.

PythonFastAPILangGraphLangChainOpenAIOllamaReact

What it does

Instead of a single large model, it runs multiple agents, each with its own role and tool set, working together. Agents break a task into steps, call tools when needed and hand off results to one another.

How it works

  • LangGraph state graph: nodes are agent/tool steps, edges are conditional transitions. The flow is state-aware; one step's output drives the next.
  • Flexible model layer: works across cloud (OpenAI) and local (Ollama) models.
  • FastAPI backend manages agent execution; a React frontend shows the process and results live.

Stack

Python · FastAPI · LangGraph · LangChain · OpenAI / Ollama · React. Containerizable with Docker.