I build AI products from prototype to production.
Engineer-turned-PM with a background in open source, developer platforms, growth, and applied AI.
My edge is connecting technical depth, product judgment, and business outcomes: understanding what is worth building, prototyping it quickly, and turning the useful parts into products people adopt and pay for.
Selected AI work
More detail in my portfolio →
How I work
Problem first. Prototype early. Measure what changes.
I start with the user and the economics, not the model. What problem is painful enough to solve? What behavior should change? What would make the product meaningfully better than the existing workflow?
Then I prototype the riskiest assumptions. A working system usually reveals more than another round of slides: interaction problems, latency, missing APIs, failure modes, cost, and where human judgment still belongs.
Once the idea survives contact with reality, I work with engineering and design to turn the useful parts into a product that can be measured, operated, and improved.
Product range
AI is my current focus, but the foundation is broader product work across growth, platforms, and developer experience.
- AI products - agentic workflows, RAG, semantic search, AI-native UX, evals, human oversight
- Developer products - APIs, MCP, documentation, onboarding, platform UX
- Growth - activation, conversion, pricing, packaging, retention, experimentation
- Business - product strategy, GTM, unit economics, positioning, 0→1 validation
Track record
- At ButterCMS, led AI, developer-experience, and activation initiatives including AI Assistant, MCP, docs consolidation, and a +40% trial-to-paid conversion lift
- At Qtravel.ai, owned the AI product roadmap for semantic search and workflow automation, including +80% booking conversion on partner integrations
- At Huuuge Games, built an M&A intelligence platform that surfaced Traffic Puzzle / Picadilla Games, later acquired for $38.9M
- At AirHelp, built automation and ML products during the company's growth from 10 to 250 people and $1M to $85M ARR
- Earlier in my career, worked on open-source infrastructure as a Google contractor
Writing
I write about the parts of AI product work that become difficult after the demo: product architecture, human oversight, agentic UX, developer workflows, reliability, and unit economics.
Start with The Agentic AI Playbook - a practical deep dive on building 0→1 agentic products from first principles.
Or browse the blog and AI-Native PM.