You raised on AI.
Now ship the version that works.
I build the architecture your AI work runs on. From day one, or after the first version stopped scaling. Production code, opinionated patterns, and the agent setup so your team and their tools stay aligned.
Greenfield
Production-ready foundation in 4 weeks. Starter codebase, core modules, Claude configuration, and deploy pipeline. From $30K.
Brownfield
4 weeks to audit a codebase that has stopped scaling, refactor one vertical slice as the reference, and plan the rest. From $20K.
Fractional Architect
Senior architecture and AI judgment on retainer. Embedded part-time, async by default. From $7K per month.
"Coding agents will write most of your code. I make sure the architecture they write into still makes sense in 12 months."
Most teams shipping with Cursor and Claude get four to six months of velocity before the patterns drift, the abstractions soften, and every new feature costs more than the last. I deliver the substrate that absorbs agent output without rotting. Either at greenfield, or after the rot has set in.
Python
TypeScript
PostgreSQL
Docker
Anthropic
Supabase
The same methodology, on my own code.
Client work sits behind NDAs. So I build in the open. Lab projects are where I run the architecture I recommend through the one test that matters: shipping it myself, on my own time, before anyone writes a check.
Barter.
A trust-focused barter marketplace. Users list items, discover matches, and negotiate trades without money changing hands.
Under the hood it is the Greenfield starter, running in production. FastAPI and Supabase on the backend, DDD with an outbox-driven event core, React and TypeScript on the frontend, and the full Claude subagent and MCP setup I ship to clients. Same patterns. Same opinions. Tested on my own dime.
More lab projects ship as the work matures. The point is not the products. The point is that the architecture I recommend is the same architecture I run.
Two builds. One retainer. Fixed scope.
Greenfield to start from day one with the right foundation. Brownfield to repair a codebase already fighting your team. Fractional Architect when the work is continuous.
Let's talk about what
you're building.
I focus on a small number of high-impact engagements. If your problem is complex and the stakes are high, we should talk.