Consulting

I help with three kinds of infrastructure problems. Each page says what I do, how I check that it worked and where I've done it before.

LLM inference

Faster LLM serving on the GPUs you already pay for

Most slow or expensive inference setups I see aren't short on hardware. They're running a default config, a batch size nobody chose, or a node shape that made sense for a different model. I find out which, fix it, and show you the numbers.

ML platforms

ML platforms your team stops thinking about

An ML platform is working when scientists ship models without filing tickets and the cloud bill makes sense. I build them on Kubernetes with Terraform and GitOps, small enough that your team can run them after I leave.

Agent sandboxing

Somewhere safe for coding agents to work

I run a fleet of coding agents against my own production systems every day. That's taught me what they need: a place to run with no credentials, a short list of things they can reach, and a way to measure whether they did the job. I build the same for other teams.

How a project usually goes

We talk through the system and what's hurting.
I measure it, so we agree on the starting number.
I make the smallest change that moves that number and show you the before and after.
Your team keeps all of it: code, dashboards and notes.

Get in touch

If something here sounds like your problem, send me a few lines about what you're running.

zac@zacp.xyz