coming next!
By joining our mailing list, calendar events will automatically appear and be updated on your calendar - no need to add them manually.
Jae-Won Chung
PhD student at University of Michigan
April 20, 2026 at 4:00 PM PT -> my local time Add to Calendar
Location: Rm 786, Davis Hall, Berkeley Watch Live on Zoom / YouTube
Energy-Efficient Machine Learning Systems poster
Abstract
Energy is becoming a critical bottleneck for scaling machine learning (ML) systems. Yet energy remains relatively less understood and under-optimized compared to traditional performance metrics like time. This talk presents principled computing systems and techniques for measuring and optimizing energy across the ML computing stack. With tooling for precise measurement, we first see that energy is a cross-layer metric that is impacted by decisions across the whole stack. Then, we introduce systems that optimize energy consumption with existing first-class metrics such as time and accuracy in mind. Finally, we discuss ongoing efforts to integrate AI datacenters and grids for tighter power and energy coordination.
Find more background information at: `ml.energy`.
Speaker Bio
Jae-Won Chung is a fifth year PhD student in Computer Science and Engineering at the University of Michigan. He works on building efficient software systems for AI workloads that view energy as a first-class systems resource that should be carefully optimized and allocated based on precise measurement and understanding. His research has been published in top venues such as SOSP, NSDI, and NeurIPS, and his open-source works have received wide recognition from academia and industry including the PyTorch Foundation, GitHub, NVIDIA, and Google.
Other Upcoming Seminars
Jerry Anunrojwong Postdoctoral Fellow at University of Washington , incoming Assistant Professor at Yale University
Title: Battery Operations in Electricity Markets: Strategic Behavior and Distortions
Yifan Liu PhD student at Georgia Institute of Technology
Title: Why EV Charging is So Unreliable? The Effects of Competition on Supplementary Service Quality