On-premise by default
Inference runs on hardware you own, inside your perimeter, not on shared cloud compute behind an API.
About
Engineering-first, founder-led, EU-domiciled.
Lucrion designs and builds private AI infrastructure for European organisations that cannot, or will not, send their data to a public cloud. We specify, procure, harden, and hand over on-premise systems engineered to a measured standard, then train your team to run them without us. The engineering credibility is the product.
What we do not do
Being explicit about scope is part of the engineering discipline. Lucrion deliberately does not take on:
- No SaaS built on your data. We build infrastructure you own, not a service that holds your data.
- No marketing-chatbot consulting. We are an infrastructure engineering firm, not a prompt agency.
- No reselling US public-cloud GPU under a European label. On-premise means on your hardware.
- No B2C products. Every engagement is with an organisation that owns the outcome.
- No crypto or Web3 work. The focus is private AI infrastructure for regulated European organisations.
Who is behind it
Lucrion is founder-led. The founder is an infrastructure and systems engineer based in Brunn am Gebirge, Austria, with a background building hardened, high-availability compute for workloads where data control and uptime are not optional. The on-premise AI playbook Lucrion sells was first built and proven on the founder's own infrastructure before it was offered to anyone else.
The team is small by design. An engagement is led by the engineer who does the build, not handed from a salesperson to a junior. That is why Lucrion runs a small number of engagements at a time rather than chasing volume.
How we are different
Open-weights, no lock-in
Open-source and open-weights throughout, so you can change models or operators without renegotiating a licence.
Compliance-aware engineering
Designed against GDPR, the EU AI Act, DORA, and NIS2 from the architecture phase, documented for external review.
Measured, not asserted
Every engagement carries acceptance criteria agreed up front, so 'it works' is a number you can check.
Reference deployment
The rig that proved the playbook
Our first reference deployment runs an associated venture's analytics workload on a hardened on-premise rig: FreeIPA and Keycloak identity, HAProxy and Keepalived high availability, NetBox inventory, MLflow, an NVIDIA GPU with DCGM monitoring, ZFS on TrueNAS, and encrypted offsite backup. It is engineered to financial-grade standards, and it is where the Lucrion playbook was first proven end to end.
It is a reference deployment running an associated venture's workload, not a customer reference. We frame it honestly because the engineering is what matters.
Austrian-domiciled · EU jurisdiction
Open positions
We are pre-launch. The founding team is small by design. Direct enquiries are welcome via the contact page. There is no recruitment funnel behind a "join us" form.
Get Started
Your data. Your infrastructure. Your control.
Start with a Readiness Assessment, a fixed-scope engagement to decide whether on-premise AI makes sense for your organisation, before any hardware is procured.