I'm a systems generalist interested in infrastructure for AI, with experience across hardware/software co-design, inference, and storage and memory systems.

I work at Meta as a Research Scientist on the AI and Systems Co-design team, where I build and optimize agentic infrastructure for personal intelligence, powering Meta's Muse. My work spans:

• Agentic infrastructure: software/hardware co-design and performance optimization for Muse models
• Memory for long-context inference: pathfinding emerging memory technologies, such as high-bandwidth flash, to scale inference capacity and cost-efficiency
• Agents for hardware co-design: agentic systems that automate software/hardware co-design

I hold a Ph.D. in Computer Sciences from the University of Wisconsin-Madison (2026), advised by Andrea and Remzi Arpaci-Dusseau in the ADvanced Systems Laboratory (ADSL), where I worked on heterogeneous storage and memory systems. I also hold B.S. degrees from the University of Michigan and Shanghai Jiao Tong University (2019).