01 Kubernetes
Clusters that scale on their own and stop waking people up.
- Kubernetes
- AKS
- EKS
- Helm
- Karpenter
Where I can help
Kubernetes, monitoring, cloud and bare metal. One engineer, from design to on-call.
What I take on
02 GitOps delivery
Every change goes through Git, reviewed and easy to undo.
- ArgoCD
- GitHub Actions
- Azure DevOps
03 Monitoring
Alerts only for what matters, so you know before your users do.
- Prometheus
- Grafana
- OpenTelemetry
- Loki
04 Cloud cost
Your bill read line by line and cut to real use. One client saved 30%.
- Terraform
- Crossplane
- KubeCost
05 Bare metal
Your own servers instead of renting the cloud, run with the same automation.
- Proxmox
- ZFS
- Ansible
06 Ostoja: an office you own →
Files, documents, archive and one login on your server, moved off Microsoft 365 in stages.
- Nextcloud
- Collabora
- Paperless
- Authentik
I'm an independent DevOps engineer based in Poland (EU), with 10 years of running production systems. I design, build and operate what your software runs on: Kubernetes clusters, delivery pipelines, monitoring, cloud accounts and bare-metal servers.
You work with one person from the first look to on-call. No account manager and no handoff: the person who designs it builds it and answers the alerts. Everything is built from code, so it can be reviewed, rebuilt and handed over.
For a small business, that means reliable IT without hiring a whole team. For an engineering team, it means an experienced extra pair of hands on Kubernetes, GitOps and cloud cost, working in CET hours.
Two I'm allowed to talk about
Fintech / B2B
Financial SaaS platform scale-out
A fintech SaaS shipped its monolith through a deploy nobody trusted. I rebuilt delivery on AKS with ArgoCD, and Git became the only way to change production.
30% Lower cloud bill
5 min Deploy time, down from 1 hour
0 Deploys with downtime
Retail / SaaS
Multi-region e-commerce failover
What if a whole region goes away? Terraform and Crossplane keep an identical standby ready, so losing a region costs minutes, not data.
< 2 min Recovery time (RTO)
2 Regions, one on standby
100% Built from code
From first look to long-term care
- 01 Audit & plan → What's genuinely urgent, each fix costed over three years.
- 02 Build & migration One goal delivered in stages, while the company keeps working.
- 03 Ongoing care Updates, monitoring and tested backups, an IT lead on call.