I ship production AI systems solo, at team speed, by directing Claude Code end to end.
I am an operations-and-automation builder who turns messy real-world business problems into production software. I did not come from a traditional CS background. I learned to build by shipping, and I build and run a private production AI platform that a real company depends on daily. Details available on request.
Open to Forward Deployed Engineer, Applied AI Engineer, and AI-native builder roles (remote US).
Claude Code (daily driver) · MCP · tool-calling LLM agents · evals · RAG
Next.js · React · TypeScript · Supabase / Postgres · Python · SQL · AWS
- ai-visibility-mcp — an MCP server that tells an AI assistant how discoverable a website is to AI search: crawler access,
llms.txt, structured data, and more. A small, self-contained example of the tool-building I do every day. - claude-infra — a cross-platform CLI that audits, retrofits, and bootstraps Claude Code project infrastructure. Scores any repo 0-100 on token-efficiency, long-term-success, and repeatability. Stdlib only, Windows-native and POSIX.
- Portfolio: dallinrowley.vercel.app — the fuller story and a work sample.
I treat AI coding tools like a team I direct: I decompose the problem, hold the architecture, and run tight quality gates (lint, type-check, tests, and an adversarial self-audit of my own code). I have caught real security issues in my own work before they shipped, because trustworthy-by-construction beats hope.