Haoyu (Jeffrey) Zhai
I’m currently a Ph.D. student at the University of Illinois Urbana-Champaign (UIUC), advised by Prof. Gang Wang in the Siebel School of Computing and Data Science. My research interests lie broadly in Trustworthy AI and LLM Agents, with a particular focus on agent security, AI safety, and usable security.
My recent research focuses on understanding and improving the security and reliability of LLM-powered agents, including red-teaming coding agents and developing safer code generation models. My previous work explored real-world security and privacy challenges using diffusion models and LLM-based agents, such as identity de-anonymization and explainable phishing detection. More broadly, I’m interested in building secure and reliable AI systems that can be safely deployed in real-world applications.
My Curriculum Vitae can be found here
News
- 04/2026: Our work ReCAP is accepted at ICML 2026!
- 09/2025: Our work Revelio is accepted at IEEE S&P 2026! See you in San Francisco!
- 09/2025: PurpCode is accepted at NeurIPS 2025!
- 07/2025: Our work PurpCode won the 2025 Amazon Nova AI Challenge ($250k prize)!
- 05/2025: Our paper SmishX is accepted at SOUPS 2025!
Publications
[ICML 2026] CAPTCHA Solving for Native GUI Agents: Automated Reasoning-Action Data Generation and Self-Corrective Training
Yuxi Chen, Haoyu Zhai, Chenkai Wang, Rui Yang, Lingming Zhang, Gang Wang, Huan Zhang
The 43rd International Conference on Machine Learning, 2026[IEEE S&P 2026] Revelio: Blurred Images Can Still Disclose Your Identity
Haoyu Zhai*, Shuo Wang*, Qingying Hao, Pirouz Naghavi, Gang Wang
The 47th IEEE Symposium on Security and Privacy, 2026[NeurIPS 2025] PurpCode: Reasoning for Safer Code Generation
Jiawei Liu*, Nirav Diwan*, Zhe Wang*, Haoyu Zhai, Xiaona Zhou, Kiet A. Nguyen, Tianjiao Yu, Muntasir Wahed, Yinlin Deng, Hadjer Benkraouda, Yuxiang Wei, Lingming Zhang, Ismini Lourentzou, Gang Wang
The 39th Annual Conference on Neural Information Processing Systems, 2025[SOUPS 2025] Can You Walk Me Through It? Explainable SMS Phishing Detection using LLM-based Agents
Yizhu Wang, Haoyu Zhai, Chenkai Wang, Qingying Hao, Nick A. Cohen, Roopa Foulger, Jonathan A. Handler, Gang Wang
The 21st Symposium on Usable Privacy and Security, 2025