Skip to content

I write software that isn't allowed to fail.

Twenty years of it: air-traffic control, radar, and neonatal life-support ventilators. These days I build AI systems that verify their own work instead of trusting it, and off the clock I cure charcuterie, hunt, and still throw the occasional karate punch. It's all more connected than it looks.

You should never trust a single actor. You trust a system built on independent verification, layered review, and evidence.

  • Projects


    Short intros and links to everything I'm building, from a C++ logger used in production to AI orchestration tooling.

    Browse projects

  • Blog


    Engineering notes, deep dives, and write-ups. Newest first.

    Read the blog

  • GitHub


    Every repo in one place: the battle-tested C++ libraries, the AI tooling, and the experiments that haven't earned a card yet.

    github.com/kjellkod