CS PhD student · UC San Diego
I work at the intersection of computer graphics, vision, and machine learning, where I build neural surrogate models to simulate the physical world, and I’ve recently become interested in multimodal post-training : )
I’m always happy to chat or explore ideas for collaboration. Feel free to email me anytime!
I'm co-advised by Ravi Ramamoorthi and Tzu-Mao Li. Previously, I was an undergraduate at Cornell, where I worked with Bharath Hariharan, Steve Marschner, and Gordon Wetzstein.
ACM Transactions on Graphics (SIGGRAPH Asia), 2025
We adopt Gaussian surfels as the primitives and build an efficient framework for differentiable light transport, inspired by classic radiosity theory, enabling efficient relighting and geometry reconstruction.
SIGGRAPH, 2024
We present a method that uses arbitrary neural network architectures as control variates with automatic differentiation to create unbiased, low-variance, and numerically stable Monte Carlo estimators for various problem setups.
SIGGRAPH Asia, 2023
Using neural field caches to reduce variance in Monte Carlo PDE solvers. Our method excels when computing is limited and enables effective bias and variance trade-off.
ICCV, 2023
Introducing a Basis Transformation Method to avoid backfilling in learning new representations for modern retrieval systems.
Research Scientist Intern
High-Fidelity Physics Research Team
Summer 2025 – June 2026
Mentors: Rohan Sawhney, David I. W. Levin, Ken Museth
Research Scientist Intern
Adobe Research
June – September 2026
Mentors: Yiwei Hu, Haitian Zheng
I enjoy ballet, rock climbing, ceramics (both hand-building and wheel throwing), and playing guitar. I used to sing in a band called Cornell Interlude.
I have served as a teaching assistant for the following courses.