Shawn (Wanxiang) Zhong
Hi, I’m Shawn (Wanxiang) Zhong, a final-year Ph.D. candidate in Computer Sciences at the University of Wisconsin–Madison, advised by Andrea and Remzi Arpaci-Dusseau in the ADSL lab.
My research explores how AI and other modern technologies are reshaping operating systems. My recent work (YoloFS) addressed the widespread corruption and leaks of user data by AI agents, where I built an agent-native filesystem that keeps data safe while preserving agent autonomy. Earlier, I worked on deterministic Linux scheduler testing (kSTEP), eBPF kernel extension reliability (DepSurf), and persistent memory filesystem performance (MadFS).
More broadly, I contribute to machine learning systems like Triton and PyTorch. I also enjoy building interactive visualizers across systems and math: Verus Explorer for system verification internals, Triton support in Compiler Explorer for IR and the optimization pipeline, xv6 filesystem visualizer for on-disk filesystem layout, JsSpim for the MIPS instruction set, and Jenn3D for 4D polytopes.
I’m graduating in August 2026 and looking for industry research and engineering opportunities. The best way to reach me is by email, happy to chat!
📚 Publications
Don’t Let AI Agents YOLO Your Files: Shifting Information and Control to Filesystems for Agent Safety and Autonomy. Preprint. [arXiv]
Shawn Wanxiang Zhong, Junxuan Liao, Jing Liu, Mai Zheng, Andrea C. Arpaci-Dusseau, and Remzi H. Arpaci-Dusseau.
kSTEP: Characterization and Controlled Testing of Linux CPU Scheduler Bugs. OSDI ’26.
Tingjia Cao, Shawn Wanxiang Zhong, Caeden Whitaker, Ke Han, Andrea Arpaci-Dusseau, and Remzi Arpaci-Dusseau.
Revealing the Unstable Foundations of eBPF-Based Kernel Extensions. EuroSys ’25. [Paper] [Code] [Dataset]
Shawn Wanxiang Zhong, Jing Liu, Andrea Arpaci-Dusseau, and Remzi Arpaci-Dusseau.
MadFS: Per-File Virtualization for Userspace Persistent Memory Filesystems. FAST ’23. [Paper] [Code]
Shawn Wanxiang Zhong, Chenhao Ye, Guanzhou Hu, Suyan Qu, Andrea Arpaci-Dusseau, Remzi Arpaci-Dusseau, and Michael Swift.