About

I started my career as a software tester. Watching how rarely software delighted the people using it — and how often it frustrated them — is what set the direction for everything after: interfaces that are friendly to humans, and moving complexity off people so a system handles it instead of them. "I moved complexity so that humans didn't have to wrestle with it. It was just handled by the system." The first real instance was in my first developer job: a Unix system that had to feed data to a mobile device through SAP's flat-file interface, one file at a time, with people manually mapping values across too many screens. I built a go-between server with its own small mapping language instead, so one composed feed reached the device rather than many.

That pattern held at every stop since. At Patient Wisdom, working for Nate Perry-Thistle as lead developer, I took a prototype built by two junior developers straight out of bootcamp to market-ready in about two months, I think. Alongside the product, I built the team — developing the processes and investing in people so they became "really excellent software engineers who practice clear thinking and excellent software disciplines." One contractor I mentored there, also fresh out of a coding bootcamp, went on to become a senior software engineer at Airbnb — or maybe, by now, a staff engineer; I'm honestly not sure which. At Chime, over three months, I led a project that took on-call — by my own account, "basically a full-time job manually repairing data errors," often more than 40 hours a week — down to a couple of hours of verification a week, by building a system that catalogued failure types and auto-applied corrections, logged for audit. And when engineers on one team grew reluctant to write end-to-end tests because a complex UI made them flaky, I built capybara_narrative, a small language for describing pages that hides the brittle field mappings inside a page-object file, so tests read clean and stayed precise. It's still public on GitHub.

What I write about now follows the same thread: what it takes to get a team working well together, and what the extra capacity AI creates should be spent on. My view, stated plainly: "a person working 60-hour work weeks should be able to go down to 30-hour work weeks and still get more done than they did before, if we do it right." The gain should also mean more presence with family, community, and each other. Some of what I write about is checkable rather than just asserted. The discipline I picked up building the team at Patient Wisdom — that a backlog item has to be written so "anybody who picked up a backlog item could execute it and pick up the context" — is, today, literally the file my AI agents load before they start work: it still asks whether an item could be "reviewed cold and handed to a parallel agent without a meeting."

Rad Nine Software is the company I founded, based in Burnaby, BC. It partners with Perry Thistle Inc., Nate Perry-Thistle's US corporation, on a 50/50 profit share — I lead technical execution and process, Nate leads business strategy. "The Perry Thistle engagement is a mutual transformational partnership where together we are building a company." It's the same relationship that started at Patient Wisdom, where I worked for Nate as lead developer years before either company existed; we stayed in touch, and now we're building this one together.

If you want to reach me, use the contact form on this site rather than an email address. I don't want my email getting scraped, so there isn't one published here.