Chicago, working with teams anywhere in the US

Platform and AI engineering for growing teams

I set up the cloud, deploy pipelines and AI tools that small and mid-size companies run on. At my day job I run 21 Kubernetes clusters in three AWS regions.

Six services for small teams. Fixed price, or monthly.

  • Cost and security audit1 week

    Waste and risk in your AWS account, fixed in Terraform.

  • Deploy pipeline setup1 to 2 weeks

    GitLab CI, Helm and Argo CD, with tests on every merge.

  • Kubernetes upgradesFixed price

    Off EKS extended support, at $0.60 a cluster-hour.

  • Shared modules and pipelinesFixed price

    Terraform modules, Helm library charts, CI components.

  • Safe AI setup2 to 3 weeks

    Company AI accounts, sign-on and a written use policy.

  • Fractional platform engineerMonthly

    Set hours a month to keep it patched and monitored.

Email me about a project

How I work

Every change goes through a merge request, passes its tests before it ships, and comes with a runbook, so your team can run it without me.

Platform work at a multi-region SaaS company. Done in my role as a platform engineer.

Runs 21 clusters from one Git repo

Argo CD hub-and-spoke on EKS in three AWS regions. Moving a live cluster onto it passed all 535 in-cluster health checks.

Moves live traffic between clusters

Weighted DNS added to the shared Helm library chart every service uses, driven by Kargo. Promotions sync in seconds, not three minutes.

Split a Terraform monorepo into modules

Over 60 resources in four regions moved into reusable modules with moved blocks, so nothing was destroyed or rebuilt.

Moved Kafka onto Amazon MSK

Managed Kafka with SCRAM sign-in, and CI runners that reach it across AWS accounts.

Projects I run at home. The same practices, on my own servers.

Runs about 30 apps like production

Every change is a merge request with an AI review, a smoke test and automatic rollback.

Keeps the servers up with an AI agent

It reads alerts, opens tickets and merges low-risk fixes. What it can reach sets its limits, not its prompt.

Signs in a household with one account

Single sign-on across the apps, secrets in Infisical, and a nightly scan that files each vulnerability as an issue.

About Ernest Coleman

I'm a platform engineer in Chicago. By day I build the platform a multi-region SaaS company ships on: AWS, Terraform, Kubernetes, Argo CD and Istio.

I consult for companies that need that kind of work but not a full-time hire. I take a few clients at a time, and each project is scoped in writing before it starts.

I don't take GovCloud work. On AI policy I write the technical controls, and your own lawyer or accountant signs off.

Questions before a first call

Who do you work with?

Small and mid-size companies that run on AWS and have no one whose job is to own the infrastructure.

How do you price?

Projects have a fixed price, agreed in writing before work starts. Ongoing work is a monthly retainer for a set number of hours.

What do we have at the end?

The code in your own repository, a runbook for each system, and a walkthrough with your team.

Will you sign an NDA?

Yes. I never name a client or show its systems without written permission.