Research Overview

Mimee Xu builds privacy-critical machine learning infrastructure. She is a Research Fellow at the Information Law Institute at NYU and earned her PhD in Computer Science from the NYU Courant Institute in 2025.

Current Directions

Her work focuses on two complementary directions:

(1) achieving exact unlearning through data-flexible training, including training that supports reversible data inclusion (as introduced in Netflix and Forget), and

(2) establishing Secure Computation as foundational infrastructure for evaluation over encrypted datasets and computations, as well as for data acquisition and model auditing over sensitive data.

Together, this work enables mechanisms for accountable oversight.

Academic Bio

As a Research Fellow at the Information Law Institute at NYU, Xu’s ongoing work on Privacy of the Mind examines how AI-mediated cognition poses a novel systemic risk, and how emerging technical capabilities could modernize privacy protection law and inform the development of new legal doctrines.

Xu earned her PhD in Computer Science from the NYU Courant Institute in 2025. Her dissertation advanced scalable secure computation techniques for dataset valuation, model evaluation, and encrypted auditing of models.

Prior to her doctorate, she applied machine learning to the engineering of large-scale production systems. Her industrial experience includes using machine learning to replace labor-intensive areas of production system management at Google, research engineering at Baidu Silicon Valley AI Labs, research and development at UnifyID, and doctoral research internships at Meta AI Research and ByteDance Applied Machine Learning.

She is also an interdisciplinary field-builder, having co-organized and grown Machine Learning for Systems — a longstanding NeurIPS workshop — for seven years, and is particularly interested in in-person coordination across technical, legal, and policy domains.