About

I am a Ph.D. candidate in Computer Science and Engineering at UC Santa Cruz, advised by Andrew R. Quinn, and expect to graduate in 2027. My research focuses on systems software for AI infrastructure: GPU runtimes and policy mechanisms, distributed-training performance diagnosis, programmable Linux runtimes, and system-level observability and control for autonomous agents.

I build research systems rather than isolated prototypes. Recent work includes SysOM-AI, a cross-layer performance-diagnosis system for production AI training; bpftime, a userspace eBPF runtime and the basis of my OSDI 2025 work on safe application extensibility; gpu_ext, which makes GPU driver and device policies programmable; and AgentSight, which correlates AI-agent activity with operating-system effects.

I also founded and lead the eunomia-bpf open-source community. Its original systems projects and educational resources have accumulated more than 10K GitHub stars and are used by external developers, researchers, and downstream integrations.

In Summer 2026, I am a Research Scientist Intern (AI Infrastructure) at ByteDance. Before my Ph.D., I worked on userspace eBPF at Imperial College London, open-source systems at PLCT Lab, static analysis at Naive Systems, and developer infrastructure at Repomono.

For a concise overview, see my CV, Google Scholar, or GitHub.