Shichong (Nio) Peng

My goal is 3D that we can truly work with, not just look at. I build 3D deep learning systems and neural renderers that turn real images into explicit 3D structure (neural point clouds and 3D Gaussians) and then predict how it moves and deforms over time. The common thread is 3D that is both photorealistic and editable: something you can edit, animate, and control, not just watch.

I'm a PhD candidate at Simon Fraser University, advised by Ke Li, and I'm finishing my PhD in 2026. Most recently I was a Research Scientist Intern at Meta Reality Labs (Codec Avatars), where I built a transformer-based decoder and a learned dynamics model that make loose clothing move naturally on full-body photoreal avatars. Earlier in my PhD I worked on neural rendering (PAPR) and on the IMLE family of generative models. I'm on the research job market for 2026.

Outside of research, I'm passionate about inner exploration and understanding ourselves and the universe, and I'm happiest out in nature. I'm always open for a chat, so feel free to reach out.