I'm a Product Manager at AWS where I help enterprises transform legacy systems into intelligent, autonomous workflows powered by agentic AI. By day, I'm turning "that's not possible" into shipped products. By night, I prototype the weird, useful agents I wish existed.
I'm obsessed with developer experience. If it takes more than 3 commands to run, it's not done. Everything I build ships with code, clear docs, and a "just works" setup.
Where the code lives
Fun Agentic Apps — Production-ready agents you can clone and run in minutes. Calendar negotiation, predictive maintenance, baby care tracking, Montessori activities. Each one solves a real problem with real code.
Agentic AI Insights — No fluff, no hype. Framework comparisons, architecture patterns, and production lessons—with code examples and actual numbers.
Agentic Playground — 20 production-ready demos across LangGraph, CrewAI, AutoGen, and AWS Bedrock. Clone → install → run. Each demo is documented, tested, and deployable.
What I care about
- Developer experience first — if it's hard to use, it's my fault
- Working code over architecture diagrams
- Honest framework evaluations — what actually works at scale
- Making enterprise AI less painful to ship
What I'm building right now
- Multi-agent orchestration patterns that don't require a PhD to understand
- RAG pipelines that actually retrieve the right stuff
- Framework benchmarks with reproducible results
- Tools that make agent debugging less miserable
🤝 Working on something interesting in the agent space? Let's talk—I love nerding out on architecture decisions and what's actually working. Also writing small checks ($25-50k) into early-stage agentic AI startups.