Available — 2 client slots, Q3
Platform engineering for the AI era.
Twenty years building the infrastructure other engineers ship on. Twelve-thousand-node fleets, multi-cluster Kubernetes, cloud bills cut close to half. Now doing it for a small number of clients directly.
What I do
Platform foundations
Terraform, EKS, GitOps — the declarative kind that still holds up when nobody is watching. Your product team ships without filing a ticket first.
AI enablement
Agents and models need what every other workload needs: identity, cost controls, observability, a way to roll back — and data pipelines that do not fail quietly. I make the platform underneath them safe to move quickly on.
Cost and reliability
Where the money goes, why the cluster wobbles, and what to do about it. Usually a two-week look before anyone commits to anything bigger.
Track record
I am Gerard Ceraso, principal at Focal Plane. Most recently engineering manager for DevOps at Buoy Software, running multi-cluster EKS through ArgoCD. Before that: a $20B ticketing marketplace moved off bare metal into Kubernetes, a 12,000-node security platform, and a long run of cloud bills brought back down. CISSP since 2010.
How I work
A call
Thirty minutes. You describe what hurts. If I am not the right person, I will say so.
A two-week look
Fixed fee. I read your Terraform, your pipelines and your bills, then hand you a written plan you own whether or not you hire me again.
Build or advise
Hands on the keyboard, or alongside your team while they do it. Whichever gets you off the critical path faster.
Writing
Nothing exists in production without being declared in code
Most of the infrastructure problems I get called about are not really infrastructure problems. They are memory problems. Somebody made a change by ...
Tell me what is breaking.
Seed to Series A with no platform team, or a scaleup whose platform team is underwater — both are the job. I read everything myself and reply within a day.