CASTIEL2: Multi-GPU AI Train the Trainer Workshop

The Train the Trainer Program is provided in conjunction with the regular AI and multi-GPU courses. This Train-the-Trainer program runs alongside the standard AI and multi-GPU computing courses. While the regular course focuses on practical skills for building and scaling AI workloads across multiple GPUs, the Train-the-Trainer track is dedicated to preparing future instructors to teach these topics effectively. 

Too few people are currently able to offer advanced AI and multi-GPU training at the level required by scientists, engineers, and PhD students who need to scale AI models efficiently or optimize distributed training workflows. Across Europe, only a limited number of HPC and AI centres provide such specialized courses. With this program, we aim to empower additional trainers and centres to deliver high-quality AI and multi-GPU training - either across Europe or within their own countries.

During this training event, participants will gain a deep understanding of modern multi-GPU and distributed AI technologies, along with practical techniques for teaching these concepts effectively. Expert instructors will guide attendees through state-of-the-art tools, frameworks, and best practices for scaling AI workloads, while also providing pedagogical strategies, teaching resources, and hands-on experience designed to prepare participants to deliver their own AI and multi-GPU training courses.

Instructors

Ashwin Mohanan (RISE)
Mats Sjoberg (CSC)
Caspar van Leeuwen (SURF)
Gyula Ujlaki (NCC Hungary)
Eleonora Bergamaschi (CINECA)
Michele Visciarelli (CINECA)
Prof. dr. Geert Jan Bex (UHASSELT)