Service 04
Infrastructure that scales quietly.
Infrastructure should be boring. If your team is thinking about the deploy pipeline, something is wrong with it. I build the pipes, the alerts and the autoscaling, then get out of the way — and I'll usually cut your cloud bill on the way through.
Start a project ↗What I build
- Cloud cost reduction
- Auditing a bill and cutting it. Idle capacity, NAT gateways, orphaned storage, environments that never sleep — the savings are almost always in the same handful of line items.
- Kubernetes & containers
- Cluster setup, deployments, autoscaling, and the manifests to run it. Also telling you when you genuinely do not need Kubernetes.
- Infrastructure as code
- Terraform modules for reproducible environments, so staging matches production and nobody clicks through a console.
- CI/CD pipelines
- Build, test and deploy automation on GitHub Actions — with preview environments and rollbacks that work under pressure.
- Observability
- Metrics, logs, traces and alerts that fire on things that matter, so you find out from a dashboard rather than a customer.
Stack
- Orchestration
- Kubernetes · Docker · Helm
- IaC
- Terraform · Pulumi · CloudFormation
- Clouds
- AWS · GCP · Azure · Vercel · Fly.io
- Pipelines
- GitHub Actions · observability · autoscaling
Whether this is a fit
A good fit if
- Your cloud bill has grown and nobody knows exactly why
- Deploys are manual, scary, or both
- You need infrastructure set up properly once, not a permanent hire
Probably not if
- You need 24/7 on-call — I build and hand over, I'm not an ops rota
- You want a migration with no downtime window and no budget for testing
- The problem is actually the application, not the infrastructure
Questions I get asked
Can you cut our cloud bill?
Usually. Idle compute, NAT gateway data processing, unattached volumes and 24/7 non-production environments account for most overspend I find. I audit first and tell you the number before you commit.
Do we need Kubernetes?
Probably not. Most teams under a certain size are better served by a managed platform. I'd rather tell you that than bill you for a cluster you'll struggle to maintain.
How do you charge?
Three ways, whichever suits the project: an hourly rate, a monthly contract for ongoing work, or milestone-based payments tied to agreed deliverables. We settle which one on the scoping call, before any work starts.
Will we be locked into you?
No — that's the point of infrastructure as code. Everything is in Terraform in your repo, documented, and handed over. You can take it to anyone.
Do you do migrations?
Yes — between clouds, or from bare servers to managed services. These need a proper testing window; I will not pretend otherwise.
Related writing
- Why Your AWS Bill Is 3× What It Should BeThe specific line items that quietly inflate an AWS bill — idle capacity, NAT gateways, forgotten storage, and cross-AZ traffic — and how to find yours in an afternoon.
- Kubernetes When You Don't Need KubernetesDeployment options form a ladder, and most teams climb two rungs past the one their workload justifies. The real cost of a cluster, and the conditions where Kubernetes actually earns it.
- Terraform vs Pulumi for Small TeamsA head-to-head on HCL versus real programming languages for infrastructure, covering state, providers, drift, the OpenTofu licence split, and a decision procedure for teams under ten people.