Join us for a 60-minute webinar ! Modern AI development increasingly depends on efficient, scalable workflows that can take full advantage of High-Performance Computing (HPC) systems. This webinar brings together three complementary perspectives on how to optimize the AI lifecycle on HPC, from model tuning to workflow orchestration and scientific application integration.

Together, these talks provide an overview of the tools, methods, and use cases that are shaping the next generation of scalable AI on HPC systems.

This presentation introduces hyperparameter optimization (HPO) as a key technique for improving machine learning and deep learning models, highlighting both the practical challenges of large-scale tuning and the opportunities offered by HPC platforms, with examples from high-energy physics.

The second talk presents itwinai, an emerging framework designed to simplify and streamline AI workflows across distributed computing environments.

The final talk focuses on the integration of domain-specific scientific use cases within itwinai, showing how such tools can support real research applications and help bridge the gap between AI infrastructure and scientific discovery.