A small, independent team that cares more about building good software than looking busy. We build and run distributed systems — device management, building automation, private cloud, data, identity, and operations — the kind of unglamorous infrastructure other businesses quietly depend on every day.
We sweat the details, tell the truth, treat people right, and keep learning.
We don't care where you went to school, whether you finished, or where you worked before. We care about who you are now and what you can do today. If you're a self-directed manager of one who can be handed a fuzzy problem and turn it into something dependable, you'll feel at home.
This role is about ensuring quality in an AI-first world. Not clicking through test plans by hand — building the systems that prove our software works, and building them with AI.
You'll automate quality assurance through and with frontier AI models: generating and maintaining test suites, hunting for the edge cases humans miss, reasoning about regressions, and catching problems before a customer ever sees them. You're a product engineer who happens to be obsessed with quality — close to the code, close to the user, and allergic to "it works on my machine."
A real engineer, not a gatekeeper. Comfortable with TypeScript, JavaScript test ecosystems, CI/CD, and writing the automation that keeps a fast-moving product honest. You know how to put frontier AI models to work on testing — and how to tell when they're wrong. Bonus if you've worked on distributed systems, where the interesting failures hide between the services.
The stack is the least interesting thing about you. What matters more: you can learn, reason, debug, and ship.
We work in small, independent teams on six-week cycles, with far fewer meetings and interruptions than you're probably used to.
We're AI-first and AI-native. We expect you to reach for frontier AI models and get everything you can out of them — that's how we build now.
We'd rather read a clear paragraph than sit through a status meeting, and ship a small finished improvement than babysit a half-built one. We value judgment and ownership — no blaming, no "not my job."
You treat quality as a design problem, not a gate at the end. You like teaching machines to catch what people miss, you take responsibility for what ships, and you're senior enough to say "this isn't ready yet" and mean it.
You think quality is someone else's job once the code is written. You only want to run someone else's test scripts. Or you're suspicious of AI instead of curious about how far it can take the work.
Make sure what we ship actually works — and build the AI-assisted systems that prove it, fast. Not perfect. Just dependable, and getting more so every cycle.
Interested? Tell us about a bug you caught (or wish you had), and how you'd build a system to catch the next one. That tells us more than a résumé.