News

• 04/2026 Our paper on benchmarking alignment of coding agents is accepted to CAIS'26.
• 02/2026 Our team PurCL is selected as one of the 10 teams in Amazon Nova AI Challenge year 2. The challenge focuses on evaluating user-centric coding agents: how coding agents effectively interact with a user and help users audit AI-generated code.
• 07/2025 🏆 Our team PurCL won 1st Place in the Amazon Nova AI Challenge. The challenge focuses on security alignment of code language models. Our team won 1st place in a challenge involving 10 top universities around the world, selected from more than 90 applicants. Code: github.com/PurCL/ASTRA.
• 06/2025 - 09/2025 I am interning at Microsoft Research in Redmond this summer. During the internship, I studied the architecture of tool designs in coding agents. Our research formulates how tool interfaces affect the robustness, exploration, and efficiency of coding agents.

Selected Publications

1. When Harmful Intent Dissolves into Technical Detail: How Safe Are Coding Agents Against Cyber Misuse?

   Xiangzhe Xu, Shiwei Feng, Guangyu Shen, Xiangyu Zhang. CAIS'2026. PDF 👐Included in OpenHands Benchmarks

2. From Poisoned to Aware: Fostering Backdoor Self-Awareness in LLMs

   Guangyu Shen, Siyuan Cheng, Xiangzhe Xu, Yuan Zhou, Hanxi Guo, Zhuo Zhang, Xiangyu Zhang. ICML'2026. PDF 🎖ICML Spotlight Paper (Top 2.2%)

3. ASTRA: Autonomous Spatial-Temporal Red-teaming for AI Software Assistants

   Xiangzhe Xu*, Guangyu Shen*, Zian Su, Siyuan Cheng, Hanxi Guo, Lu Yan, Xuan Chen, Jiasheng Jiang, Xiaolong Jin, Chengpeng Wang, Zhuo Zhang, Xiangyu Zhang. Technical Report. PDF 🎖1st Place Solution in Amazon Nova AI Challenge

4. TAI3: Testing Agent Integrity in Interpreting User Intent

   Shiwei Feng*, Xiangzhe Xu*, Xuan Chen, Kaiyuan Zhang, Syed Yusuf Ahmed, Zian Su, Mingwei Zheng, Xiangyu Zhang. NeurIPS'2025. PDF 🎖3,000 USD Bug Bounty by Finding Bugs in Alexa+

5. ProSec: Fortifying Code LLMs with Proactive Security Alignment

   Xiangzhe Xu*, Zian Su*, Jinyao Guo, Kaiyuan Zhang, Zhenting Wang, Xiangyu Zhang. ICML'2025. PDF

6. GenNm: Symbol Preference Aware Generative Models for Recovering Variable Names from Stripped Binary

   Xiangzhe Xu, Zhuo Zhang, Zian Su, Ziyang Huang, Shiwei Feng, Yapeng Ye, Nan Jiang, Danning Xie, Siyuan Cheng, Lin Tan, Xiangyu Zhang. NDSS'2025. PDF

7. RepoAudit: An Autonomous LLM-Agent for Repository-Level Code Auditing

   Jinyao Guo, Chengpeng Wang, Xiangzhe Xu, Zian Su, Xiangyu Zhang. ICML'2025. PDF

8. CodeArt: Better Code Models by Attention Regularization When Symbols Are Lacking

   Zian Su, Xiangzhe Xu, Ziyang Huang, Zhuo Zhang, Yapeng Ye, Jianjun Huang, Xiangyu Zhang. FSE'2024. PDF

9. ReSym: Harnessing LLMs to Recover Variable and Data Structure Symbols from Stripped Binaries

   Danning Xie, Zhuo Zhang, Nan Jiang, Xiangzhe Xu, Lin Tan, and Xiangyu Zhang. CCS'2024. PDF 🎖ACM SIGSAC Distinguished Paper Award

10. LLMDFA: Analyzing Dataflow in Code with Large Language Models

    Chengpeng Wang, Wuqi Zhang, Zian Su, Xiangzhe Xu, Xiaoheng Xie, Xiangyu Zhang. NeurIPS'2024. PDF

11. ProRec: Source Code Foundation Models are Transferable Binary Analysis Knowledge Bases

    Zian Su, Xiangzhe Xu, Ziyang Huang, Kaiyuan Zhang, Xiangyu Zhang. NeurIPS'2024. PDF

12. ROCAS: Root Cause Analysis of Autonomous Driving Accidents via Cyber-Physical Co-mutation

    Shiwei Feng, Yapeng Ye, Qingkai Shi, Zhiyuan Cheng, Xiangzhe Xu, Siyuan Cheng, Hongjun Choi, Xiangyu Zhang. ASE'2024. PDF 🎖ACM SIGSOFT Distinguished Paper Award

13. ParDiff: Practical Static Differential Analysis of Network Protocol Parsers

    Mingwei Zheng, Qingkai Shi, Xuwei Liu, Xiangzhe Xu, Le Yu, Congyu Liu, Guannan Wei, Xiangyu Zhang. OOPSLA'2024. PDF 🎖ACM SIGPLAN Distinguished Paper Award

14. DiEmph: Improving Binary Code Similarity Transformer Models by Semantics-Driven Instruction Deemphasis

    Xiangzhe Xu, Shiwei Feng, Yapeng Ye, Guangyu Shen, Zian Su, Siyuan Cheng, Guanhong Tao, Qingkai Shi, Zhuo Zhang, and Xiangyu Zhang. ISSTA'2023. PDF

15. PEM: Representing Binary Program Semantics for Similarity Analysis via A Probabilistic Execution Model

    Xiangzhe Xu*, Zhou Xuan*, Shiwei Feng, Siyuan Cheng, Yapeng Ye, Qingkai Shi, Guanhong Tao, Le Yu, Zhuo Zhang, Xiangyu Zhang. FSE'2023. PDF

16. StateLifter: Extracting Protocol Format as State Machine via Controlled Static Loop Analysis

    Qingkai Shi, Xiangzhe Xu, Xiangyu Zhang. USENIX Security'2023. PDF

17. ARCTURUS: Full Coverage Binary Similarity Analysis with Reachability-guided Emulation

    Anshunkang Zhou, Yikun Hu, Xiangzhe Xu, Charles Zhang. TOSEM'2023. PDF

18. CSLED: Automatic Generation and Validation of Instruction Encoders and Decoders

    Xiangzhe Xu, Jinhua Wu, Yuting Wang*, Zhenguo Yin and Pengfei Li. CAV'2021. PDF

19. CompCertELF: Verified Separate Compilation of C Programs into ELF Object Files

    Yuting Wang, Xiangzhe Xu, Pierre Wilke, Zhong Shao. OOPSLA'2020. PDF

20. CPC: Automatically Classifying and Propagating Natural Language Comments via Program Analysis

    Juan Zhai, Xiangzhe Xu, Yu Shi, Guanhong Tao, Minxue Pan, Shiqing Ma, Lei Xu, Weifeng Zhang, Lin Tan, Xiangyu Zhang. ICSE'2020. PDF

Experience

• Jun. 2025 - Sep. 2025 Research intern at Microsoft Research. Advisors: Qianhui Wu, Hamidreza Saghir, Marc-Alexandre Cote, Tong Wang, Kiran Lakkaraju, Michael Albada. Microsoft Research.
• Apr. 2021 - Aug. 2021 Research assistant on binary program analysis. Advisor: Charles Zhang. HKUST.
• Sep. 2020 - Feb. 2021 Research assistant on program verification. Advisor: Yuting Wang. SJTU.
• May 2020 - Aug. 2020 Intern on automatic differentiation. Advisor: Hao Chen. ByteDance AI Lab.
• Dec. 2019 - Mar. 2020 Research assistant on program analysis. Advisor: Xiangyu Zhang. Purdue University.
• Jul. 2019 - Oct. 2019 Research intern on program verification. Advisor: Zhong Shao. Yale University.
• Jul. 2018 - Jun. 2019 Research intern on program analysis. Advisors: Minxue Pan, Juan Zhai. Nanjing University.

Awards

• Amazon Nova AI Challenge Research Grant ($250,000 + $1M computation resource), 2026
• Bug Bounty of Alexa+ ($3,000), 2025
• 1st Place in Amazon Nova AI Challenge ($250,000), 2025
• Amazon Trusted AI Challenge Research Grant ($250,000 + $1M computation resource), 2024
• 1st Place in AutoDriving CTF at DEFCON30, from 110 global teams, 2022

Invited Talks & Lectures

• Scaling auditing expertise via neural-symbolic multi-agent system, Feb 2026, Amazon
• Intelligence engineering in the era of vibe coding, Oct 2025, Purdue University
• Building trustworthy AI coding systems through agentic red-teaming, July 2025, Amazon; May 2025, TrustNLP
• Scaling security expertise with AI-driven systems, Apr 2025, RIT; Mar 2025, Microsoft
• Harnessing domain expertise to elevate post-training data quality, Mar 2025, Meta
• Understanding programs when symbols are lacking, Nov 2024, UMass Amherst
• An agentic red-teaming framework, Nov 2024, Amazon
• Inference time scaling for code reasoning task, Oct 2024, Purdue University
• Incorporating program analysis insights to code models, Apr 2024, UMass Amherst
• Introduction to code language models, Nov 2023, Purdue University

Services

Other Services

• The 42nd International Conference on Software Engineering (ICSE'20) Track Scheduling co-Chair
• Faculty Search Representative in Purdue Computer Science Graduate Student Association (2023-2024)