About me

I’m currently a post-doc researcher at NVIDIA Research. Starting in the Summer of 2025, I will be an assistant professor in the Department of Computer Science at Purdue University. I received my Ph.D. degree from the EECS Department of MIT advised by Srini Devadas and earned my B.S. degree in Mathematics from Tsinghua University.

I am broadly interested in provable trustworthy computation. My work focuses on developing theoretical frameworks to quantitatively model emerging societal concerns, such as privacy, adversarial robustness, and copyright protection, in cutting-edge data processing tasks. On the practical side, I design algorithms for black-box trustworthiness analysis, enabling efficient and automatic constructions of provably trust-preserving solutions by treating the target data processing function as a black box, without requiring additional assumptions or knowledge. Some of my recent works study

1. data-specific indistinguishability between a target model and a set of safe models for provable data-influence based trustworthy guarantees, including memorization mitigation, defense against backdoor attacks and copyright protection for black-box machine learning algorithms (ICML 2025);
2. semantic and rigorous measures which are accessible to a general audience with board expressibility of security and privacy concerns (CRYPTO 2023, IEEE S&P Magazine 2025);
3. end-to-end automatic privacy proof and privacy-preserving technology (IEEE S&P 2025) that accommodates cutting-edge advancements in data science and machine learning (ICML 2020, PODS 2022, CCS 2024);
4. fundamental relationship between security & privacy, the minimal utility/efficiency overhead (CCS 2023, IEEE S&P 2023) and (Byzantine/Adversarial) robustness (SODA 2024) for algorithm co-design and simutaneous improvement.

My earlier works encompass a range of related topics including Byzantine consensus (TCC 2020a, TCC 2020b), and proof of work (TCC 2017) in applied cryptography, and constrained sampling theory in signal processing (IEEE-TSP 2023a, IEEE-TSP 2023b).

I was a recipient of Mathwork Fellowship (2021-2023) and Tsinghua Future Scholar Fellowship (2015-2017) . My research has also been supported and funded by DSTA, Singapore, Captical One and Cisco.

I am always looking for students who are passionate about developing useful theory to guide practical solutions. If you’re interested in working with me, feel free to send me your CV.

Current PhD Students

• Justin He (2025-, UCSD, B.S. in Mathematics + Computer Science)
• Wendi Wang (2025-, UIUC, Zhejiang University, B.S. in Computer Engineering + Mathematics)
• Zhouyi Zhang (2025-, Peking University, B.S. in Mathematics)
• Zhizhen Yuan (2024-, Tsinghua University, B.E. in Computer Science, co-advised with Ninghui Li)