Advanced Training
MINERVA training activities are dedicated to both academic and industrial users, and aimed at bridging the knowledge gap on the efficient use of HPC resources that is typically experienced by the AI community.
The training material repository is accessible HERE.
Tutorial: AI Agents – GCPR 2025
This tutorial highlights two essential frameworks for developing robust, multi-step agents: LangGraph and the Model Context Protocol (MCP). For additional resources, please refer to the companion repository.
HPC for AI Research – ICIAP 2025
This tutorial focuses on the computational aspects of training and deploying large-scale AI models using high-performance computing (HPC). Topics include distributed training, multi-GPU and multi-node acceleration, mixed-precision optimization, efficient data […]
Training Calendar
Upcoming
What you can expect in the next months
Understanding User Needs for Better AI Courses
We will identify academic and industry networks to assess their HPC systems proficiency, use cases, and computational needs. Our approach aims to create courses for a broad audience while focusing on prevalent use cases. We’ll survey users associated with HPC providers and expand to include academic communities and industries, using existing networks and establishing new connections. Our surveys will be disseminated through a website and direct contact with community representatives. We aim to maintain a comprehensive understanding of skill distribution, use cases, and HPC experience, and we will monitor its evolution during the project.
HPC Courses for AI Model Training
We aim to develop and deliver HPC courses to communities and end-users, catering to both entry-level and experienced users, focusing on distributed training or inference of ML/AI models in HPC environments. The courses will cover topics including Large Language Models, Multimodal Models, and Vision-only models, emphasizing large-scale training and data management around open foundation models. Course materials, including slides and working code samples tailored for EuroHPC supercomputers, will be regularly updated. We will deliver the courses through in-person and streaming events, targeting AI community users and entities that interact with such users.
Ensuring Quality and Sharing Training Resources
We will continuously assess the quality of developed courses by monitoring user satisfaction levels and tracking skill increments within the community, aligning with approaches for community monitoring. We are responsible for devising corrective actions to enhance the quality level and suggesting adjustments to the course organization. A website and repositories for hosting all training materials will be included in the MINERVA portal. We will collaborate with WP6 for course promotion through social media and dissemination channels available to the Consortium.
Hackathons: Engaging the AI Community
The MINERVA project envisions hackathons as part of its strategy to strengthen the Community Hub and engage with the ML/AI community. The Cross Institutional Team (CIT) will organize events like workshops and hackathons around open foundation models research and development, as well as their transfer to downstream applications. Digital information, such as blog and social media posts, will highlight these events. These initiatives are designed to maximize the project’s outreach and drive engagement with the broader AI community.
Further training
Have a look on what is available from our fellow EuroHPC JU-funded projects on the HPC in Europe Portal.