// Karlsruhe · since 2026
Infrastructure that stops waking you up at night.
Cloud migration, Kubernetes, and the unglamorous work of making a system somebody else built survive its next year. The engineers who answer your call also write the code.
Fixed scope, two weeks, findings you keep either way.
- deploy frequency
- per day
- 0.3
- →
- 4
- commit → prod
- lead time
- 2 d
- →
- 3 h
- change failure
- rate
- 24 %
- →
- 9 %
- time to restore
- p50
- 95 min
- →
- 22 min
The four DORA metrics, before and after a typical engagement.
Figures are illustrative, not client data.
What we do
- 01
DevOps & platform engineering
Pipelines that finish in minutes and environments that match. We rebuild the path from commit to production and leave it documented.
- 02
Cloud security review
An adversarial read of your cloud estate — IAM, network, secrets, supply chain — ranked by what an attacker reaches first.
- 03
Reliability & incident response
SLOs your team believes, alerts that mean a human should act, and runbooks written before the outage rather than during it.
- 04
Operations you can hand back
We run it while you hire, then document and transfer it. Every engagement is designed to end.
Stack
What we work in
- Public clouds
- AWS
- Google Cloud
- Azure
Where the estate lives, and where most of the cost and blast radius sits.
- Containers & orchestration
- Kubernetes
- Helm
- Docker
Kubernetes built and operated, not just advised on.
- CI/CD
- GitHub Actions
- GitLab CI
- Jenkins
- Argo CD
The path from commit to production, and how long it takes to walk it.
- Infrastructure & configuration
- Terraform
- Ansible
- Puppet
- Vault
So the environment can be rebuilt, reviewed and rolled back like code.
- On-premise & hardware
- Proxmox
- KVM
- VMware vSphere
- Bare-metal provisioning
Owned hardware is a first-class target, in both directions — cloud to datacenter and back.
- Monitoring, tracing, logging
- OpenTelemetry
- Prometheus
- Grafana
- Elastic
Knowing something is wrong before a customer tells you — and which service caused it.
- Automation
- Python
- Ruby
- TypeScript
The glue: tooling, migrations and the jobs nobody wants to run by hand.
- [ CLIENT_01 ]
- [ CLIENT_02 ]
- [ CLIENT_03 ]
- [ CLIENT_04 ]
- [ CLIENT_05 ]
Find out what breaks first.
Two weeks, fixed scope. A written verdict on what will fail, ordered by how badly, and what it costs to prevent. Yours to keep either way.