Industry Panel: Power & Energy in Production ML

Panel

Bhargava Gopireddy NVIDIA Cooper Elsworth Google Arun Tejus Raghunath Rajan Meta/MLCommons Jae-Won Chung (Moderator) University of Michigan

Abstract

Join industry experts from NVIDIA, Google, and Meta as they discuss the real-world challenges and solutions for power and energy management in production ML systems. Learn about infrastructure planning, operational strategies, and what every ML practitioner should know about power constraints.

Panel Outline

How is the power bottleneck manifesting in infra planning and operation?
What are your short- and long-term projections for power/energy issues?
If you can make one change magically in the ML stack, what would it be?
What would be one thing you would like ML researchers and practitioners to take away from today's discussion?