RI-SCALE_D3.1 – AI Systems and Models Specification and Roadmap

The Data Exploitation Platform (DEP) enables Research Infrastructures (RIs) to scale their AI applications across large-scale computing infrastructure, such as on cloud and High Performance Computing (HPC) systems, and enables scientists to train and/or run AI models at scale with RI scientific data. Work Package (WP) 3 plays a central role in enabling these capabilities by providing the technical solutions needed to integrate AI functionalities in the DEP. This deliverable outlines the technical specifications of the AI solutions proposed in WP3 for the DEP, detailing their main features, planned developments and integrations. The document also presents user stories that illustrate scenarios of accessing the DEP for different kinds of users. These stories highlight the DEP access mechanisms and the role of the software solutions in the DEP. The AI applications and their compute and data requirements are also defined in this document. These requirements and the user stories guide the modular architecture of WP3, which enables flexible and customizable definition of workflows for DEP users. Finally, the document proposes the creation of testbeds, which form the methodological basis for realizing the technical implementations in the DEP.