The Czech EOSC Node contributes to the EOSC Federation by providing Czech national e-infrastructure and services, thus providing researchers with access to advanced computational resources, secure sensitive data environments, and research data repositories.
Its advanced capabilities include LLM-as-a-Service, Galaxy, Kubernetes, Jupyter, SensitiveCloud and federated file sync and share, while providing an onboarding framework for national repositories and strategic multi-node use cases with other EOSC Nodes.
The entire Czech ecosystem will be connected to the EOSC Federation through federated AAI built on the Perun platform. The Node’s specialised services, including the SensitiveCloud Trusted Research Environment and a dedicated Service Incubator for emerging services to be incorporated in the Node, help support FAIR data practices across the EOSC Federation.
Key objective: To showcase the work of Czech researchers and institutions by connecting their research outputs and services to the EOSC Federation.
Science areas: Specialised services dedicated for all science areas, including FAIR data management, Trusted Research Environments and Virtual Research Environments, and AI services.
EOSC Node Czechia expands EOSC technical capacity through secure sensitive-data environments, federated AI services, Galaxy workflows, and curated national repositories supporting cross-border research.
FAIR data
The EOSC CZ Node provides FAIR research data through national repositories including MBDB, GENASIS, LINDAT/CLARIAH-CZ and other domain-specific repositories, supported by the National Metadata Directory, repository onboarding framework, Data Stewardship Wizard for machine-actionable DMPs, and a Service Incubator for onboarding emerging FAIR-compliant services.
Scientific use cases
Cross-Node federated LLMs
This use case establishes a cross-Node federated Large Language Models (LLMs) inference service for EOSC users, providing access to LLMs through a unified web interface and an OpenAI-compatible API.
The service enables researchers and research communities to use LLMs interactively, integrate them into scientific workflows, notebooks, platforms and domain portals, and develop domain-specific AI assistants.
Federation increases the available capacity, model diversity and resilience of the service, while reducing fragmentation and making better use of distributed European GPU and AI resources. It also enables participating providers to expose their resources through a common access layer and improve their utilisation.
The EGI EOSC Node provides the federation layer, while participating Nodes and providers—currently EOSC Node Czechia and SZTAKI from Hungary—contribute distributed LLM capacity.
AI-enhanced federated Galaxy workflows
This use case builds upon the established European Galaxy network, integrating the Czech usegalaxy.cz instance with the wider multi-node network (including usegalaxy.eu operated by the NFDI node).
As a pioneer in the second wave of EOSC Nodes, the Czech Node will introduce a new feature: the integration of Galaxy workflow systems with the Federated LLM-as-a-Service (LLMaaS). Building on an existing successful prototype between the .cz and .eu instances, this project will mature the integration to production quality to provide advanced analytical pipelines for the ELIXIR and NFDI communities.
This allows researchers to utilize Large Language Models directly within their scientific pipelines for tasks such as automated data annotation, metadata enrichment, and code generation for custom analytical steps. By integrating LLMs into Galaxy, researchers can automate complex interpretive tasks that previously required manual intervention, significantly accelerating the research lifecycle.
Other use cases
- SensitiveCloud for MCVAL (multi-centric validation of AI models for cancer screening)
- Federated File Sync and Share services
- Federated analysis of Sensitive Data in TRE with the EOSC Node Switzerland
