What we do
Three things, in the order they have to happen: move the infrastructure, build the agents in, and let the architecture carry both.
The move, not the destination
We move your systems from Azure, AWS or Google to European infrastructure. Scaleway first; UpCloud in Copenhagen, STACKIT or Netic when the ground has to be Danish or German. Not as a new platform purchase, but as a scoped engineering job with a known end.
Everything is written as code. Network, Kubernetes, databases, secrets, logging and access live in Terraform in your own repository, so the estate can be recreated in an afternoon, and so you can move again. Security and the supervisory controls are built into that code, not into a document beside it.
DORA asks for an exit plan from your cloud provider. We deliver it as something that can be run: a workload that demonstrably comes up at a second European provider, with the hours recorded. And we make the cryptography replaceable, so the EU’s roadmap for post-quantum cryptography does not become yet another move.
Portable architecture · The rehearsed move
In an ordinary engagement we do not promise that the deliverable can be moved to another cloud. We build what you ask for, on the provider you have chosen, and use the managed services that make sense there.
If you choose Portable architecture, we guarantee the opposite: everything we deliver can be moved to another European provider, or back again. We do that by building on open source that exists at every provider, and by never placing a proprietary service in the critical path without a documented substitute.
“Cloud agnostic” is a phrase everyone uses and nobody believes, because it is a promise. Ours is a test. The rehearsed move: we bring one of your workloads up at a second European provider, record the hours and whatever had to be rewritten, and repeat it once a year. DORA asks for exit plans to be tested. We make the test the product.
It costs some convenience. The hyperscalers’ best services are exactly the ones that tie you down. We show you that bill before you choose.
Intelligence on your own ground, not at the hyperscalers
Nomad Intelligence is our name for the AI‑native when it does not live at the hyperscalers: open models running on European cloud or locally, and autonomous agents that do work inside the business, in case handling, in operations, in customer contact, not only in the development department. On infrastructure you control, with governance you can show the supervisor and that passes the AI Act.
The agents move the code, the architects keep it safe. We use agents to modernise old systems faster than people can type, and we keep them inside boundaries people have drawn. It is the only way 20-year-old code arrives safely on a new estate.
No agent is better than the data it stands on. We build the data foundation first: event streams as the source of truth, Data Act requirements built in, evaluations and guardrails from day one, so you can answer what the agent did, and why.
Why it holds
Domain-driven design gives us the unit of the move: a bounded context with its own data and its own APIs that can be moved without the rest standing still. It is also the unit the agents can work safely within.
Event-driven architecture gives us the data backbone: streams of events that can be read from both sides of a move, and that make parallel running possible. API design gives us the interfaces that hold when something on the other side changes provider.
The architecture is not a product we sell. It is the reason the first two things work when the auditor and the traffic arrive.
How the three fit together
The move gives your systems ground you control. The agents are the reason it is worth doing now. The architecture is the reason it works.