ETHUX Logo
ETHUX
ETHUX · Ethical User Experience

Self-hosted AI,end to end.

I'm Gijsbert. An NVIDIA-certified AI engineer in the Netherlands. I deploy LLMs on your GPUs, build the cluster they run on, and write the apps that use them. Self-taught, hands-on the whole way.

Start a conversation See recent work
ethux@rig01:~
ethux@rig01:~$ kubectl get pods -n llm
NAME STATUS AGE
vllm-mistral-medium-3-5 Running 2d
vllm-minimax-m2-5 Running 5d
gliner-daemon Running 12d
ethux@rig01:~$ nvidia-smi --query-gpu=memory.used,memory.total
23104 MiB, 24576 MiB
21888 MiB, 24576 MiB
22912 MiB, 24576 MiB
... 5 more
ethux@rig01:~$
The principle

Not your model,
not your knowledge.

If you don't own the model, you don't own the knowledge.

Infrastructure

The setup behind
the work.

0×
RTX 3090 GPUs
0
GB VRAM
0
Certifications
0+
Years self-hosting

Recent work

Open source I maintain, client work I'm engaged on, and the side experiments I keep building.

Client engagement

Freelance Forward Deployment Engineer at Gaide

Building AI deployments and PII-aware LLM tooling for Dutch enterprises.

gaide.nl

Side experiments & hackathons

Where I prototype the wild stuff

Vibe AR

Hackathon

Vibe-coding in mixed reality on a Meta Quest 3. Built in 24h at the Mistral AI Worldwide Hackathon Paris. Top-7 finalist.

github.com/ethux/Vibe-AR

MCP Infra Agent

Hackathon

MCP server that creates VMs and deploys Docker containers from any GitHub repo, in natural language. Mistral AI MCP Hackathon, Paris.

Out in the wild

Mistral AI Ambassador · NVIDIA GTC ′25 & ′26 · KubeCon · TalosCon · Mistral hackathon finalist

Want AI infra that stays yours?

Tell me what you're trying to do. I'll tell you honestly whether I can help, what it'll take, and what I'd build.

Start the conversation