Summer School 2026

Sustainsys Summer School focuses on sustainability and efficiency in data systems and AI/ML infrastructure, bringing together students and researchers. Participants will explore the design of  systems across the stack - from hardware and systems software to machine learning workload.

Tentative Schedule

9:00 am: Doors Open.

9:30 am: Opening Remarks by organizers: Prof. Oana Balmau, Prof. Semih Salihoglu.

9:45 am: Keynote talk: Prof. Anastasia Ailamaki.

10:45 am: Break.

11:00 am: Hands-on session 1, Increasing Sustainability in Database Servers: Prof. Tilmann Rabl.

12:30 pm: Lunch break (on your own).

2:00 pm: Hands-on session 2, Measuring energy efficiency in ML pipelines with Code Carbon: Prof. Oana Balmau, Olivier Michaud.

4:00 pm: Closing Remarks by organizers: Prof. Oana Balmau, Prof. Semih Salihoglu.

4:15 pm: Social outing (optional) – Walk up Mt Royal guided by Shubham Vashisth. 

Keynote Speaker:

Confirmed Speakers:

SESSION DETAILS:

Keynote Talk: Building Principles Data Systems on Shifting Topologies, Prof. Anastasia Ailamaki

The rapid integration of AI into data management has sparked a "gold rush" for performance, often resulting in complex, black-box systems that treat hardware as an infinite, homogeneous resource. However, beneath this AI revolution lies a harsher reality: the "Memory Wall" has evolved into a "Topology Wall." Modern scale-up infrastructures—characterized by disaggregated memory (CXL), tiered NVMe arrays, and heterogeneous chiplet-based compute—no longer offer uniform performance guarantees. In this era of radical hardware diversity, we are at risk of building brittle systems that trade foundational engineering principles for opaque, heuristic-driven optimization.

To achieve true sustainability and performance in future data systems we need principled, hardware-grounded adaptation. I will present a paradigm shift from "hardware-oblivious" engines to systems that treat the interconnect as a first-class, schedulable resource. By leveraging causal analytical models and throughput-guided, adaptive mechanisms, we can build data engines that are natively aware of their own resource bottlenecks. I will share lessons learned from building systems that use AI to inform policy, not to replace logic, demonstrating that principled data management remains our most potent tool for outlasting the current AI hype cycle. By re-centering our design philosophy on topology, data movement, and functional isolation we draw a viable path to systems that are both performance-efficient and ecologically sustainable.

Bio: Anastasia Ailamaki is a Professor of Computer and Communication Sciences at the École Polytechnique Fédérale de Lausanne (EPFL), a visiting researcher at Salesforce, and the co-founder and Chair of the Board of Directors of RAW Labs SA, a Swiss company developing systems to analyze heterogeneous big data from multiple sources efficiently. She earned a Ph.D. in Computer Science from the University of Wisconsin-Madison in 2000. She has received the 2019 ACM SIGMOD Edgar F. Codd Innovations Award and the 2020 VLDB Women in Database Research Award. She is also the recipient of an ERC Consolidator Award (2013), the Finmeccanica endowed chair from the Computer Science Department at Carnegie Mellon (2007), a European Young Investigator Award from the European Science Foundation (2007), an Alfred P. Sloan Research Fellowship (2005), an NSF CAREER award (2002), twelve best-paper awards and three Test-of-Time prizes at international scientific conferences. She has received the 2018 Nemitsas Prize in Computer Science from the President of Cyprus and the 2021 ARGO Innovation Award from the President of the Hellenic Republic. She is an ACM fellow, an IEEE fellow, a member of the Academia Europaea and the US National Academy of Engineering, and has served as an elected member of the Swiss, the Belgian, the Greek, and the Cypriot National Research Councils.

Hands-on session 1: Increasing Sustainability in Database Servers, Prof. Tilmann Rabl 

Physical limitations are rapidly bringing hardware efficiency improvements to a halt. At the same time, the AI boom is demanding

enormous increases in compute capacity. In the search for profit, industry has mostly given up on former goals of carbon neutrality in a global arms race on AI. In this pivotal moment, research is needed to clear up the clouds hiding the true economic, ecologic, and societal costs of current IT trends to open alternative paths for sustainable computing.

In this talk, we will discuss current IT trends from an ecological perspective. We will analyze different measures of efficiency of data systems and methods to improve it. Incorporating estimations on hardware and power production carbon intensity, we will estimate ecological impact of hardware and review implications on data system development.

Bio: Tilmann Rabl is a Professor for Data Engineering Systems at the Digital Engineering Faculty of the University of Potsdam and the Hasso Plattner Institute. Tilmann received his doctoral degree at the University of Passau. He was a postdoctoral researcher at the University of Toronto and the Technical University of Berlin. His current research focuses on efficiency of database and ML systems, hardware efficient data processing, benchmarking, and sustainability.

Hands-on session 2: Measuring energy efficiency in ML pipelines with Code Carbon, Prof. Oana Balmau, Olivier Michaud 

This session introduces students to practical energy measurement for machine learning using CodeCarbon. As AI workloads scale, their energy and carbon footprints are becoming significant. Improving energy efficiency is critical for sustainable AI and for reducing operational costs in large-scale deployments. Participants will learn how to instrument ML workloads, interpret energy metrics, and reason about trade-offs between performance and efficiency. The session is hands-on: students are encouraged to bring their laptops to follow along, run experiments, and analyze the energy behavior of real ML workloads.

Bio: Oana Balmau is an Assistant Professor in the School of Computer Science at McGill University, where she leads the DISCS Lab. She is a part of MLCommons, where she co-founded MLPerf Storage, an open-source benchmark for storage on ML workloads. Her research focuses on storage systems and data management, with an emphasis on ML, data science, and edge computing workloads. She completed her PhD in Computer Science at the University of Sydney, advised by Prof. Willy Zwaenepoel. Before her PhD, Oana earned her Bachelors and Masters degrees in Computer Science from EPFL. 

Bio: Olivier Michaud is a computer science master's student at McGill University, advised by Profs. Oana Balmau and Bettina Kemme. His main interests are hardware-software co-design, concurrency, and distributed systems. He is currently working on understanding and improving the energy efficiency of GPU accelerated workloads.

PARTNER UNIVERSITIES: