Most of us have had that moment where we set up a machine or a dev environment “just this once,” promise ourselves we’ll document it later, and then… never do. Now imagine doing that across thousands of servers, multiple data centres, and a bunch of very latency-sensitive systems. Terrifying, right?

In this talk, I’ll share how we tackled this problem at Optiver by building our own Infrastructure as Code platform — for physical, bare-metal machines that absolutely refuse to behave like the cloud. Along the way, we learned a lot about declarative design, reconciliation loops, standards, Python tooling, and also, human nature.

You don’t need to work in trading or infrastructure (or know what a picosecond is!) to follow along. If you’re curious about building systems you can trust, reducing manual chaos, or just want to hear an honest story of how code can help tame messy realities, I hope you’ll take something valuable away.