AI engineer & product builder
I lead teams that ship LLM applications. After hours, I build my own products.
I write about what I learn from both: when you need an agent at all, how agents keep track of long tasks, and how to tune prompts against a metric instead of by feel.
Four products built after hours. All are live.
Generates sets of badges that share one art direction, so they look like they belong together.
Turns the countries and cities you visit into illustrated stamps, travel goals, and a score to share. For iPhone.
Pressure-tests a SaaS idea before you spend months building it. The free trial needs an account and a card.
Matches your profile to a job ad, then writes a cover letter that fits only that role.
Essays published here and elsewhere, with the latest pieces visible first.
A practical pipeline for turning expert judgment into reusable agents and skills that survive real work.
Why prompt engineering by intuition stops working in production, and how DSPy plus MIPROv2 let you optimize prompts against explicit metrics.
A practical definition of AI agents, what actually counts as agentic, and when simpler LLM systems are the better engineering choice.
Showing 3 of 6 pieces
Leading teams that build production LLM applications.