CLASTIQ.ai
The CLASTIQ AI Platform combines NVIDIA GPU hardware with Clastiq Enterprise in one installed, supported system. Teams deploy open models, run notebooks and fine-tune on shared GPUs, while administrators govern access, capacity and usage from a common control plane.
One box, every team
Each project has its own members, roles, storage, secrets and quotas. Teams use the shared system without administering the hardware themselves, with one solution covering hardware, platform and support.
GPU sharing and model serving
Allocate whole GPUs or fractions as small as a quarter card, queue work and lend idle capacity within defined limits. Choose vLLM, SGLang, llama.cpp or TensorRT-LLM per deployment.
A gateway for your applications
Serve models behind one OpenAI-compatible endpoint with project-scoped API keys and token quotas. Use the playground to test text, images and documents before connecting an application.
Build around the model
Managed notebooks, fine-tuning, custom containers and a marketplace of vector databases and AI tools support the complete workflow. Guardrails, single sign-on and audit logs give administrators oversight.
Visibility without guesswork
Monitor GPU utilisation, memory, latency, request rates and tokens alongside workload logs and events. Track each project against its allocated resources.
Hardware that fits the workload
The brief describes desk-side DGX Station and eight-GPU H200, B200 and B300 configurations. TIADA validates the selected configuration, model requirements and support scope before quoting.
Highlights
- On-premises and air-gap ready for strict data sovereignty
- One console for GPUs, bare metal, virtual machines, and Kubernetes
- FinOps engine: budgets, quotas, burn-rate monitoring, and metering
- Governed agent environment for multi-agent AI workflows
Who it is for
SMEs building a private AI capability, enterprise teams sharing GPU infrastructure, and organisations that need local control over models, prompts and datasets.
From evaluation to deployment
- Size the hardware against your model, memory, storage and concurrency needs.
- Connect identity and create projects with GPU shares, quotas and keys.
- Deploy a model, test it in the playground and connect applications through the gateway.
Deployment considerations
Model fit, throughput and availability are validated for the chosen hardware. Local and air-gapped operation requires the appropriate configuration; review any external model downloads or optional provider connections. CLASTIQ is a Cognition AI and Technology Innovation SPC product; RAQI is its UAE reseller and deployment partner.
Related pages
Work with RAQI
- Buying a product? Purchases and installation go through our channel partner, TIADA Solutions.
- Have a question? Call our AI voice agent on +971 56 100 4203.
- Not sure what fits? Book a 30-minute discovery call with our CEO.