About
Production AI engineer and consultant
Focusing on AI-native engineering practices, AI agent architecture, evals, observability, and the codebase changes that make AI useful instead of noisy.
Fifteen years of production software, most of it in rooms where “mostly works” wasn't good enough. Thus shaping my perspective on AI.
My background
My background
I've spent most of my career building software under real constraints: deadlines, regulated environments, production incidents, team growth, and the usual mismatch between how systems are supposed to work and how they actually behave.
Then
I co-taught a cohort of the Metis Ruby on Rails bootcamp — a 12-week intensive that turned out production-ready engineers. It's where I learned to compress hard-won engineering intuition into something a builder could actually use, not just admire.
I helped grow the engineering team from 0 to 50+ and shipped FDA-cleared medical device software (510(k) BK231004) — the kind of environment where "it mostly works" isn't a real answer.
Time as a Staff Engineer at Shopify, working at a scale where small decisions compound fast.
A lot of AI advice is either theoretically smart or practically battle-tested. I care about the second category.
Now
Today I consult with teams shipping production AI while keeping my hands in the work as a senior engineer at August Health and teaching what I learn in public.
I stay hands-on as a senior engineer, building August Intelligence — the AI platform for better care and operations in senior living communities.
My AI-native engineering patterns, MCP servers, and skills are open-sourced on GitHub, including claude-code-workflows — a public collection of patterns built from daily use. I also teach what I learn on YouTube.
Most of that experience now takes the shape of an agent-augmented Fractional CTO engagement — CTO-level architecture, technical direction, and shipped code, extended by agents rather than replaced by them. The judgment stays human. The agents extend the reach.
Services & Specialties
Every engagement pulls from the same handful of things I've spent the most time on. Some show up as code, some as architecture decisions, some as what I teach — but they all trace back to the same fifteen years of production work.
If you're past the demo stage and hitting real constraints, let's talk. 30 minutes. No pitch deck.