I sit at the intersection of AI engineering and enterprise adoption.

Most people can build an AI prototype. Few can navigate the organizational complexity to make it stick. I do both — and that's the rare combination that creates real impact.

• Build: AI agents with Claude, LangGraph, FastAPI, MongoDB
• Deploy: bring agents into enterprise environments (ERP, legacy systems, change management)
• Share: open-source tools and skills for the AI builder community

Enable Anthropic's Agent Skills in any Python agent

Lets LLM-powered agents autonomously discover and load modular skill packages — a token-efficient alternative to bloated system prompts. Published on PyPI, compatible with LangChain and framework-free deployments.

pip install skillkit

Curated library of ready-to-use Agent Skills

Battle-tested skills for startup founders and technical teams building with Claude and other AI agents. Think of them as superpowers for your AI.

AI-powered assistant for skydiving instructors

Serverless backend for AIR Coach: an AI agent for parachuting instructor training. Built with LangGraph + Gemini + MongoDB, deployed on AWS Lambda. A real-world example of AI in a specialized, safety-critical domain.

Dynamic multi-agent backend with LangGraph + FastAPI

Production-ready backend for creating and managing AI agents dynamically. Features automatic agent selection, stateful conversations, tool integration and S3-backed document context.