Infrastructure that stops being the bottleneck.
Deploys that take half an hour. Alerts everyone has muted. A cloud bill that grows faster than your traffic. I fix that for European B2B tech companies — in weeks, not quarters.
In IT since 2005 · DevOps and SRE since 2016
Commit to production — 2.5 min
Numbers from production, not from a brochure.
Every figure is a before-and-after from a real engagement.
Ten things companies hire me for.
Grouped by what they change: how you build, how you run it, how you prove it is safe, and where you want a second opinion.
Build
03Cloud & Infrastructure
AWS design, Terraform & Terragrunt, multi-account, cost control.
CI/CD & Automation
GitLab CI, GitHub Actions, ArgoCD — commit to prod in minutes.
Kubernetes & Containers
Cluster design, Helm charts, multi-cluster ops, autoscaling.
Operate
03Monitoring & Observability
Prometheus, Grafana, Loki, ELK — know before your users do.
K8s Cost Optimisation
Cluster analysis with ready-to-merge PRs. Paid from the savings.
Privacy-First Managed SRE
Your isolated EU instance — hosted, patched, backed up, watched.
Secure
03Security & Resilience
Hardened baselines, secrets, disaster recovery, SLOs, runbooks.
NIS2 Readiness
Gap assessment, technical remediation, audit-ready checklist.
DNS & IaC Drift Monitor
Live DNS against your Terraform state. Alerts on drift and expiry.
Advise
01Consulting & Support
Architecture reviews, team enablement, post-mortems, SRE retainers.
Fixed price, two weeks. You get a written assessment with priorities and effort estimates — useful even if you stop there.
Scoped and quoted before we start. Migrations, platform builds, pipeline rebuilds, observability from scratch.
Monthly, an agreed number of days. Ongoing SRE capacity, on-call support and someone who already knows your setup.
For cost work only. You pay a share of what the cluster actually stops costing you. If nothing is saved, nothing is owed.
Building infrastructure engineers trust.
I work with European B2B tech companies to build infrastructure that is secure, observable and built to scale — from zero to production.
Based in Helsinki, trading as a Finnish toiminimi, available across the EU and EEA.
Long enough to have made most of these mistakes once already.
How the work actually gets done.
The tools change from one client to the next. These six do not.
Everything as code
Infrastructure, networks and policy live in Git and get reviewed like application code. Nothing is clicked into existence in a console.
Terraform · Terragrunt · Ansible
Git is the source of truth
The cluster follows the repository, never the other way round. A rollback is a revert, and anyone can read what production is supposed to look like.
Kubernetes · Helm · ArgoCD
You find out before your users
Metrics, logs and traces on every service, with alerts tied to what users actually feel rather than to raw CPU graphs.
Prometheus · Grafana · Loki · tracing
Capacity follows load
Nodes and workloads scale to real traffic, and right-sizing is a continuous job — not an annual panic when the bill arrives.
Karpenter · autoscaling · cost review
Secrets never sit in a repo
Credentials come from one place, live briefly, and can be revoked everywhere at once. Access is granted by role, not by memory.
Vault · managed secrets · SSO
Recoverable by design
A backup that has never been restored is a rumour. Restores are rehearsed on a schedule and the runbook is written before the incident.
Tested restores · DR runbooks
Start with thirty minutes.
No deck, no discovery questionnaire. Tell me what hurts, and I will tell you what I would fix first, roughly what it takes — and whether you actually need outside help for it.
You keep everything
Code lives in your repositories, infrastructure in your accounts. No proprietary tooling, no lock-in, and runbooks stay behind when I leave.
NDA before anything sensitive
Mutual NDA signed before access is discussed. Credentials go through your secret manager — never email, never chat.
With your team, not around it
Your engineers keep owning the platform. I work in your review process and hand over as I go, so nothing depends on me staying.
Finnish company, EU-only infrastructure
Registered toiminimi in Helsinki, Y-tunnus 3608757-4, invoiced in EUR with Finnish VAT. Everything I host runs in Germany and Finland — no data, backups or logs leave the EU, and none are planned to.