I've been writing software for a living since 2008. I started with research software and e-commerce integrations, then moved to remote product work, managed an engineering team, and built developer tooling used across a whole company.
These days I work on AI platforms: the product and backend code around LLMs, and the agent workflows engineers use to ship with them. I also help teams work out where AI helps in their process and where it gets in the way.
The work I do most, whether I'm writing the code or leading the people who write it.
I build the software around LLMs: API integrations, agents and tool calls, MCP servers, and the backend and UI that turn a model into a product people can depend on.
Splitting systems into sensible parts, designing APIs and integrations, and planning migrations. I write RFCs and prefer designs a team can still maintain in five years.
JavaScript has been my main language for most of my career. I've written Node.js services, React applications and a shared UI kit that other teams build on.
CI/CD, shared tooling, stack standardization and automated code review. I use AI agents in my own daily work and help other teams set up similar workflows.
I've managed a team, run planning, mentored engineers and interviewed candidates. Much of the job is splitting big projects into pieces people can finish and keeping everyone in the loop.
AI / LLM platformsTypeScript / JavaScriptReact & web platformsNode.jsGraphQLAPIs & integrationsCI/CDGitHub automationArchitecture & system design
Most of my career has been in web platforms. Over time the work moved from writing code toward architecture, team leadership and, lately, AI platforms.
Years of applied research on distributed and geospatial systems, and recent certifications for the tools I use every day.
Code and architecture are tools. What counts is that the right thing ships and keeps working after release.
I use AI agents every day. They make good engineers faster, but review, ownership and judgment stay with people.
I pick simple designs, clear ownership and proven tools. Fashionable architecture usually costs more than it gives back.
I still write code and review pull requests. It's hard to lead technical work once you lose touch with the codebase.
I can join as an engineer who writes code, or work at the architecture and leadership level. Usually it ends up being a mix of both.
If you're working on an AI product, want to change how your team builds software, or are stuck on an architecture problem, send me a message.