I’m an Assistant Professor of Business Administration at Harvard Business School and affiliate faculty in Computer Science in the SEAS. As PI of the SAFR AI  research () lab I’m focused on building more private, fair, and reliable algorithms. Prior to Harvard, I received my PhD from UPenn, math degree from Harvard, and co-founded a  data startup.

Some research I’m proud of includes some of the first work on fairness in bandits, fairness gerrymandering which was incorporated into IBM’s AI Fairness Toolkit, state of the art bounds for adaptive data analysis, and some of the first work on machine unlearning. 

Recent News:

• New preprint: “Machine Unlearning Fails to Remove Data Poisoning Attacks.”
• New preprint: “Pandora’s White-Box: Precise Training Data Detection and Extraction in LLMs.” blog.  
• 4 papers accepted to ICML GenLaw ’24 Workshop!
• 2 papers accepted to ICML 2024! 
• Machine Unlearning interviews with Working Knowledge & Axios Science 2/22/24
• We wrote the first survey on privacy issues in LLMs! 12/10/23
• Paper accepted at NEURIPS Workshop on Socially Responsible Language Models 12/16/23
• Talk at Microsoft Research New England on Privacy in LLMs! 11/27/23
• New preprints online “In-Context Unlearning: Language Models are Few Shot Unlearners” [twitter thread] & “Black-box Training Data Identification in GANs via Detector Networks” 11/1/23
• Paper “Model-based Perturbation Attacks Against Language Models” published at EMNLP 2023! 12/7/23
• Invited Keynote Address on Responsible AI at the Annual eBay AI Summit! 7/13/23