Principal engineer & AI platform builder in Toronto. Twenty-five years in software, most of it now spent on the systems around the code: platform boundaries, review loops, agent workflows, and the team habits that keep AI-generated software from turning into entropy β and leading the teams who run them.
Website Β· Field Notes Β· The Lab Β· LinkedIn Β· X Β· Instagram
- AI-assisted engineering β building software when a model is in the loop for every commit.
- The new SDLC β review as the main engineering act, with automated second-opinion gates.
- Developer tooling β MCP servers as facades, local-first review systems, and the coding-agent harness I build with.
- Judgement is the job now β AI removed the boilerplate effort; taste and review are what's left.
- Build the model a map β stop making agents rediscover everything at the start of every session.
- Postman was a workaround β the generic API client was a crutch; describe the app you want and build the specific one.
More at matthewpurdon.me/notes. There's an RSS feed and an llms.txt if you're an agent.