about

I build the software that connects physical machines to the systems that run a factory floor: device drivers, message-passing services, and the tools operators and engineers use to see what those services are doing. Most of it comes down to the same problem: two things need to talk, and something has to make sure that conversation is reliable when a socket drops or a board fails a check.

What I like about it is the physicality. A bug isn't abstract. It's a board sitting in a machine that won't move, and the fix has to work on hardware that can't be paused to debug.

I've mentored a junior developer through their first production service, and I take code review seriously in both directions: as a way to catch problems before the floor does, and as the fastest way I know to bring someone new up to speed on a codebase.

I'm part of the on-call rotation for the automation services I've built, which means the honest failure paths in this site's project write-ups aren't hypothetical. I'm the one who gets paged if they're wrong.

I use Claude Code daily, for this site and for production work. The place it earns its keep isn't generating code I couldn't write. It's collapsing the distance between deciding to do something and having it done: reading through an unfamiliar service before touching it, drafting the tedious half of a migration, catching the failure case I described but didn't handle. I still read every line before it ships.