cat profile image

Hi, I’m Billy Kong. I work on platform engineering and Kubernetes, and lately on a question that keeps getting more urgent: how do you let an AI agent operate infrastructure without trusting it?

This blog is where I write up what I build and what it teaches me, including the parts that didn’t work the first time. The code behind the posts is on GitHub, with its tests.

Current project: safe-ops-agent-k8s, a reference implementation of an AI ops agent on Kubernetes that can investigate and propose changes, but can’t change anything without passing a policy gate and a human.

If you spot a mistake, or something here is useful to you, open an issue.

Billy Kong