About

I'm Zac Peterson. I've worked on infrastructure since 2019, when I joined a cloud consultancy as a cloud engineer. I left as its managed-services practice lead, running the SRE and DevOps team that looked after clients' environments on AWS, Azure and Google Cloud.

Then came senior and staff roles at a digital marketing company, a consumer apparel company, a global energy company and, since 2025, a healthcare company, where I work on AI and ML infrastructure. The details changed every time. The constants were Kubernetes, Terraform and a cloud bill somebody wanted smaller: 60% smaller at the apparel company, 40% at the energy company.

Over the past year most of my own time has gone into GPUs and coding agents. I rented a node with eight H200s to serve GLM-5.1 for my own coding agent. I wrote a Kubernetes operator that has run more than 38,000 coding-agent benchmark attempts. And I run CI for a product of mine on a five-node BuildKit fleet in my bare-metal k3s cluster.

How I work: I measure before I change anything, I prefer the smallest change that moves the number, and I leave behind code and notes your team can run without me.

I live in San Francisco. Email is the easiest way to reach me: zac@zacp.xyz.