DocumentationIntroduction
About AI as a Service
CERIT-SC, a core component of e-INFRA CZ, is developing an on-premise AI platform that provides researchers with secure, high-performance, and interoperable tools.
Our infrastructure features cutting-edge NVIDIA DGX H100 and DGX B200 systems, enabling efficient large-scale model training and inference. It hosts open large language and generative models (e.g., DeepSeek R1, Qwen3-Coder-480B, Gemma3, and GPT-OSS-120B), accessible via an Open WebUI interface or OpenAI-compatible APIs.
Beyond conversational AI, the ecosystem supports AI-augmented workflows through JupyterHub, featuring preconfigured Notebook Intelligence plugins for local model interaction. Additionally, it includes the DeepSite web application for generating websites, visuals, and presentations, as well as an AI-powered documentation portal with contextual assistance, among other features.
Key Features
- Secure, on‑premise LLM & generative‑AI platform – data never leaves the trusted environment
- Supports privacy‑sensitive research – compliance with institutional and legal requirements
What the Platform Provides
| Category | Highlights |
|---|---|
| Compute | NVIDIA DGX‑H100 (Hopper) and DGX‑B200 (Bergamo). Petaflop‑class GPU performance. |
| Models | DeepSeek R1 (LLM), Qwen‑3‑Coder‑480B (code‑focused LLM), Gemma‑3, GPT‑OSS‑120B (open‑source GPT). Additional community models can be added on request. |
| Access | Open WebUI – interactive chat & playground |
| External usage | OpenAI‑compatible REST API - use existing openai client code, LangChain, or other integrations without modification |
| Workflow Integration | JupyterHub with pre‑installed Notebook Intelligence plugins (local model calls, autocompletion, code generation), DeepSite – AI‑driven website, visual, and presentation generation |
How to Get Started
| Step | Action | Links |
|---|---|---|
| Explore the chat UI | Test models instantly via the web interface. | WebUI chat |
| Call the API | Use the OpenAI‑compatible endpoint to integrate AI into your scripts, pipelines, or services. | Using AI models – API docs |
| Launch a Jupyter notebook | Open JupyterHub, start a notebook, and enable the Notebook Intelligence extensions for in‑notebook model calls. | JupyterHub integration |
| Generate content with DeepSite | Automatically create websites, graphics, or slide decks from prompts. | DeepSite (vibe‑coding) |
| Get Contextual Help | The documentation portal itself features an AI chatbot that can answer questions about the platform. | Embedded chatbot |
Read more details on our e-INFRA Blog at https://blog.e-infra.cz/
Last updated on
