Praneeth Karimireddy
Assistant Professor of Computer Science and Electrical and Computer Engineering
Education
- Doctoral Degree, École Polytechnique Fédérale de Lausanne
- Master's Degree, Indian Institute of Technology-Delhi
- Bachelor's Degree, Indian Institute of Technology-Delhi
Biography
Dr. Sai Praneeth Karimireddy is an Assistant Professor at USC in the Thomas Lord Department of Computer Science, and by courtesy in the Ming Hsieh Department of Electrical and Computer Engineering. Before this, he was an SNSF Postdoctoral Fellow working with Mike I. Jordan at UC Berkeley, and obtained his PhD at EPFL advised by Martin Jaggi. His work has previously been deployed across industry at Meta, Google, Open AI, and Owkin.
He is interested in principled approaches to understanding real-world AI. During his PhD, he developed efficient and robust algorithms for distributed and federated learning. However, engaging with practitioners revealed that incentive alignment is often the biggest bottleneck for real-world multi-agent collaborations. He spent his postdoc learning game theory and mechanism design to be able to address such questions. Now, he combines insights from optimization, statistics, and economics to design, evaluate, and improve trustworthy AI systems.
Selected awards he has been recognized by include:
- Amazon Center on Secure & Trusted ML award
- Capitol One Fellowship
- SNSF Mobility Fellowship
- Patrick Denantes Memorial Prize for the best thesis in computer science
- Chorafas Foundation Prize for exceptional applied research
Research Summary
He is currently interested in principled approaches to defining, understanding, and improving AI behavior. Key questions his research addresses include:
- AI Safety & Privacy: What does it mean for complex AI systems to be safe or private? How can these properties be audited, measured, and improved?
- LLM Evaluation: Machine learning has moved beyond simple i.i.d. train/test splits. How should the metrics from various benchmarks be interpreted? More generally, how can LLM behavior be accurately measured and improved?
- AI Ecosystems: How can the field understand and shape the complex "machine behavior" arising from networks of interacting humans and AI agents? How can the health of such ecosystems be ensured through the design of correct incentives?
Awards
- 2022 SNSF Mobility Fellowship
- 2022 Patrick Denantes Memorial Trust Patrick Denantes Memorial Best Thesis Prize
- 2022 EPFL Thesis Distinction Award
- 2021 Dimitris N. Chorafas Foundation Chorafas Foundation Prize
- Thomas Lord Department of Computer Science
- Ming Hsieh Department of Electrical and Computer Engineering
- SAL 327
- Henry Salvatori Computer Science Center
- 941 Bloom Walk, Los Angeles, CA 90089
- karimire@usc.edu



