Six areas where I do the most useful work. Most engagements combine a few of them — the goal is always the same: fewer manual steps between a commit and production, and fewer surprises after it.
Clusters that survive contact with production — on bare metal with Talos, or managed on EKS and AKS.
One source of truth for every environment, so releases stop depending on who is at the keyboard.
Right-sized infrastructure on AWS, Azure, and Hetzner — backed by load-test data, not guesswork.
Know what broke before the customer tells you, and know how to get it back.
Infrastructure for sites with no internet, strict audits, and no room for improvisation.
Getting agents past the demo and onto the same release train as everything else.
A platform from scratch — cluster, networking, storage, delivery pipeline, and the runbooks to operate it.
Move what already runs: Swarm to Kubernetes, VMs to EKS/AKS, or a cloud bill to something defensible.
A focused audit of an existing setup — reliability gaps, cost leaks, supply-chain risk — with a prioritized plan.
Describe the setup and what hurts about it. If it isn't something I should take on, I'll say so and point you somewhere better.