01
AI strategy & consulting
Strategy, advisory, use-case discovery and a clear roadmap to value.
- AI strategy for your business
- Use-case prioritisation
- ROI & roadmap
← Back to home
Services
Services
A complete, end-to-end offer to design, deploy, run and grow private AI inside your company.
02
On-premise infrastructure
Hardware advisory and purchase across brands, installation, deployment and scaling.
- Hardware advisory & buying
- Installation & deployment
- Growth & scaling
03
Operation & support
We keep your AI running: maintenance, management, monitoring and MLOps.
- Maintenance & management
- Monitoring & MLOps
- SLAs & support
04
AI applications
Final applications built on your private models and integrated into your tools.
- Assistants & copilots
- RAG over your data
- Automation & agents
05
Data
The fuel for AI: data engineering, preparation and data governance.
- Data engineering
- Data preparation
- Data governance
06
Governance & compliance
AI governance, GDPR and EU AI Act advisory, risk management and audits.
- AI governance
- GDPR & AI Act advisory
- Risk & audit
07
Training program
We upskill your teams so AI becomes a capability you own.
- Team training
- Hands-on workshops
- Adoption enablement
08
Custom models
We adapt and fine-tune open-source models to your domain and knowledge.
- Fine-tuning & specialization
- Evaluation & benchmarking
- Quantization & optimization
09
AI security
We protect your models, data and access end to end — even fully air-gapped.
- Hardening & isolation (air-gap)
- Access control & encryption
- Red-teaming & security testing
Models & Hardware
Many models. Many sizes. The right hardware.
We are model- and vendor-agnostic. We pick the open-source model, the parameter size and the hardware that best fit each use case, budget and privacy need — and we tell you which pairs with which.
Open-source models & parameter sizes
We work with multiple families and versions, choosing the parameter count that fits your accuracy and hardware budget.
- Llama 3.1 / 3.3 8B · 70B · 405B
- Mistral / Mixtral 7B · 8x7B · 8x22B
- Qwen 2.5 7B · 14B · 32B · 72B
- DeepSeek V3 · R1 · Coder
- Gemma 2 9B · 27B
- Phi 3.5 mini · small · medium
- Falcon 7B · 40B · 180B
Hardware — brands, models & sizes
From a single workstation GPU to a multi-node cluster. We advise, source, install and scale across vendors.
-
NVIDIA
RTX 4090 · L4 · L40S · A100 · H100 · H200
-
AMD
Instinct MI210 · MI300X
-
Dell · HPE · Lenovo
GPU servers & workstations
-
Supermicro · Intel
GPU nodes & accelerators
Models & tools for agents
Function-calling and tool-use models for autonomous agents, copilots and automation.
Hermes (function calling) Clawbot Qwen-Agent tool-use & RAG pipelines
Model ↔ hardware recommendations
We map it both ways: tell us your goal and we recommend the model + hardware, or tell us your hardware and we recommend the best model and size.
| Profile | Recommended model | Recommended hardware |
|---|---|---|
| Small team / pilot | 7–8B (Llama 3.1 8B, Mistral 7B) | 1× 24GB GPU (RTX 4090 / L4) |
| Department / production | 32–70B quantized (Qwen 32B, Llama 70B) | 2× 48–80GB GPU (L40S / A100) |
| Enterprise / high performance | 405B · Mixtral 8x22B · DeepSeek V3 | Multi-GPU cluster (H100 / H200) |
| Autonomous agents | Hermes · Qwen-Agent (tool calling) | A100 / H100 + vector store |
Not sure where to start? We benchmark options on your real use case before you buy a thing. See all services →
How we work
A clear path from idea to a private AI that you own and operate.
- 1
Advisory
We assess your needs, data and goals.
- 2
Strategy & design
Use cases, model and architecture choice.
- 3
Hardware & sourcing
We recommend and procure the right hardware.
- 4
Install & deploy
On-premise setup, models and applications live.
- 5
Training
We enable your teams to use and own it.
- 6
Maintenance & growth
We operate, monitor and scale with you.
Contact
Let's talk about your AI project
Tell us your goal and we will get back to you with a clear next step — wherever you are in Europe.