Evian McKeown

My work spans the full stack: web applications in ASP.NET Core and Django, the APIs and data models beneath them, and the algorithms that govern how quickly they respond. I am drawn to problems in which performance and correctness are both binding requirements rather than competing preferences.

In practice this has produced a DTMF decoder with an end-to-end latency below 50 ms, a live journey planner whose RAPTOR router answers in under 0.1 s and is about thirty times faster than a Dijkstra baseline, and a telemetry plugin with over 140,000 downloads. The planner began as a university project that I led with a team of three, and I have since rebuilt it and run it on my own; the decoder was released publicly under GPL v3; and several of my fixes have been merged upstream into a GNOME Shell extension with 2,100 stars.

I also take responsibility for the less visible engineering work: automated testing, continuous integration, and deploying and securing my own services on Linux, with sandboxed systemd units, SELinux, bot protection and rate limits.

Several of the projects above began as hobbies rather than assignments: the telemetry plugin came out of sim racing, and the decoder out of wanting to use a Casio Databank watch as a dialler. I also shoot 35 mm film on an Olympus OM-1, which accounts for the photographs below.