I build AI agents and LLM features that run in production, not demos.

I've built a multi-agent backend with FastAPI and LangChain that routes each request through retrieval, critique and synthesis agents, streams responses and runs evaluation loops, plus a research agent with swappable LLM providers. I work with the Claude and OpenAI APIs, LangChain, LlamaIndex and Hugging Face, and have contributed fixes to LlamaIndex and LiteLLM.

You get an agent that automates a real workflow in your business, connected to your tools and data, an API you can call from your app with streaming, a small test set so you can check quality before launch, and the source code in your repo.

The price is a starting point for an agent with up to 3 tools or integrations. Message me with the workflow you want to automate and I'll confirm scope.