Your product shipped and the cluster runs. But the bill climbs for reasons nobody can explain, and releases happen on Tuesdays because Friday feels like tempting fate. Nothing is broken enough to stop the business — which is why it never gets fixed.

I’m Arty, a freelance cloud and platform engineer building DevOps tooling since 2019, across music streaming, molecular diagnostics, live sports betting and online education. CKA, CKS, AWS Solutions Architect (Advanced Networking, Security), Google Cloud Architect, Microsoft DevOps Expert.


Where I help Link to heading

Cloud spend that outgrew the traffic. Oversized requests, orphaned volumes, NAT charges on traffic that never needed to leave the VPC. I measure first, then fix the biggest line item — and tell you which findings are rounding errors.

Releases that feel like a gamble. GitOps with FluxCD: repo layouts that stay navigable past forty apps, dependency ordering that waits for CRDs instead of racing them, tenant isolation enforced by the API server.

Kubernetes nobody’s confident in. Karpenter consolidation that repacks nodes without waking anyone, disruption budgets that respect working hours, alerts that correlate with user impact instead of narrating normal behaviour.

Security that holds up to an audit. Least-privilege IAM that survives a cluster rebuild, Kyverno policy-as-code packaged as one readable values file, Prowler scanning — plus the triage that turns 1,600 findings into the ten worth doing this week.

A second cloud, or a move between them. AWS, GCP and Azure. Multi-account governance, and honest advice on which workloads should stay put.


How we work Link to heading

Scoped project — fixed deliverables, fixed end date. Architecture review, cost audit, migration, security assessment. You get a written plan, not a deck.

Retainer — an engineer on call for reviews and steady improvement, without a full-time hire.

Hourly — one specific problem. Sometimes that’s all it takes.

Embedded — inside your team for longer efforts.


Stack Link to heading

Clouds — AWS, Google Cloud, Azure

Kubernetes — EKS, GKE, AKS, Karpenter, Kyverno, Helm

GitOps & IaC — FluxCD, Terraform, OpenTofu, Terragrunt

CI/CD — GitHub Actions, Azure Pipelines, GitLab CI

Observability — Prometheus, Grafana, Loki, CloudWatch

Security — Vault, External Secrets, Prowler, Trivy, cosign

Most of this shows up on the blog, including the problems that cost me an afternoon.


Let’s talk Link to heading

Thirty minutes, free, no pitch. Tell me what’s bothering you and I’ll tell you what I’d look at first. If you don’t need me, I’ll say so.