CV

A PDF version of my CV is available here.

Research Interest

My research covers software systems and machine learning, with the goal of enhancing the reliability of large-scale systems. To this end, I leverage and optimize LLMs/Agents with testing, reasoning, and verification techniques. To date, my research has detected 800+ critical bugs for ML systems, compilers, operating systems, browsers, databases, and smart contracts, including 40+ CVEs and 0-day smart contract exploits worth over $100K.

• System Reliability for ML. Synthesizing diverse & high-quality tensor programs across inference, training, and optimization components in ML systems, along with test oracles.
• ML for System Reliability.
  • LLM for Vulnerability Detection. Designing LLM-driven workflows to synthesize tests, test generators, and static analyzers for large-scale software systems. Building scalable multi-agent PoC validators to confirm LLM-reported bugs in million-line codebases. Evaluating LLMs’ code reasoning via test generation.
  • LLM for Formal Verification. Training and evaluating LLMs to generate verification proofs for code.

I am honored to be supported by the Capital One Ph.D. Fellowship and selected as an Anthropic AI Security Fellow.

Education

• University of Illinois at Urbana-Champaign, Aug. 2022 - Present
  • Ph.D. Student in Computer Science, advised by Prof. Lingming Zhang
• Nanjing University, Sept. 2018 - July 2022
  • B.Sc. in Computer Science and Technology, graduated with honors
  • Enrolled in Elite Class, GPA 91.2/100, rank 1/24

Publications

* denotes joint first authors

Awards

• Capital One Ph.D. Fellowship, 2025
• Travel Grant for ASPLOS 2025, 2025
• OpenAI Researcher Access Program ($5,000), 2024
• SIGSOFT CAPS Travel Grant for ESEC/FSE 2022, 2022
• China National Scholarship (Top 0.2%), 2020
• Special Scholarship for Undergraduates in Basic Science, Nanjing University (1/24), 2021
• Elite Program First-class Scholarship, Nanjing University, 2019

Professional Experience

• Research Fellow, AI Security, Anthropic, May 2026 - Aug. 2026
  • Hosted by Zygimantas Straznickas. Selected as an Anthropic AI Security Fellow
• Research Intern, Systems Research Group, Microsoft Research, May 2025 - Aug. 2025
  • Hosted by Shan Lu. Topic: Automated Proof Completion for Real-World Verified Systems
• Research Intern, Systems Research Group, Microsoft Research, May 2024 - Aug. 2024
  • Hosted by Shan Lu. Topic: AutoVerus – Automated Proof Generation for Rust Code
• SWE Intern, Project Starline, Google, May 2023 - Aug. 2023
  • Hosted by Srinivas Kaza and Lukas Murmann. Topic: JAXGL – Integrate High Performance Graphics Primitives into ML
• Research Assistant, PL/FM/SE Group, UIUC, May 2021 - Present
  • Advised by Prof. Lingming Zhang. Topic: Intersection of machine learning and software systems
• Research Assistant, SPAR Group, Nanjing University, Aug. 2020 - Apr. 2021
  • Advised by Prof. Yanyan Jiang. Topic: Testing

Talks

• KNighter: Transforming Static Analysis with LLM-Synthesized Checkers, Mar. 2026
  • CS 6501 Spring 2026 (AI-Driven Systems), University of Virginia
• Agent-Based LLMs for System Verification: AutoVerus and VeruSAGE, Mar. 2026
  • Large-Scale Data & Systems (LSDS) Group, Imperial College London
• VeruSAGE: A Study of Agent-Based Verification for Rust Systems, Jan. 2026
  • Chinese University of Hong Kong
• KernelGPT and KNighter: LLMs for Kernel Fuzzing and Static Analysis, Nov. 2025
  • COMP 790-199 Fall 2025, University of North Carolina at Chapel Hill
• AutoVerus: Automated Proof Generation for Rust Code, Sept. 2025
  • University of Virginia
• Next-Generation Fuzzing: Leveraging LLMs for Deep Learning and Compiler Systems, Nov. 2024
  • CS 6158 Fall 2024, Cornell University
• Fuzzing Automatic Differentiation in Deep-Learning Libraries, May 2023
  • Advanced Software Technologies Lab, ETH Zurich
• Fuzzing Deep-Learning Libraries via Automated Relational API Inference, Sept. 2022
  • Software Engineering Retreat, University of Illinois at Urbana-Champaign
• Free Lunch for Testing: Fuzzing Deep-Learning Libraries from Open Source, May 2022
  • iSE Symposium, Nanjing University

Service

• Program Committee / Reviewer: TOSEM, EMSE, ICLR’25
• Artifact Evaluation Committee: PLDI’24, ISSTA’24

Teaching Experience

• Teaching Assistant, University of Illinois at Urbana-Champaign, Aug. 2025 - Dec. 2025
  • CS 427: Software Engineering I
• Teaching Assistant, Nanjing University, Sept. 2021 - Jul. 2022
  • Problem Solving, a core course for the students in the elite program

Skill Stack

• Common: Python, C, TypeScript, Docker, Vim, Git, SQL, LaTeX, libFuzzer
• Machine Learning: PyTorch, TensorFlow, JAX