Hexu Zhao 赵和旭

Email: hz3496 [at] nyu [dot] edu

[Google Scholar] [GitHub] [LinkedIn] [X/Twitter]
About Research About Me (as I see it)

Since 2026, I have been on leave from my Ph.D., helping build a stealth world model lab in San Francisco. I drive ML systems optimization across our lab, accelerating training and inference for diverse workloads on clusters spanning thousands of GPUs.

I am a Ph.D. student of Computer Science at NYU Courant, advised by Prof. Jinyang Li and Prof. Aurojit Panda. My research focuses on Machine Learning Systems, especially Systems for Gaussian Splatting. My PhD research series—Grendel, CLM, and Gaian—develops a systems stack that scales 3DGS reconstruction from a single GPU to city-scale scenes across hundreds of GPUs.
Previously, I obtained my bachelor's degree in 2023 from Honored Yao Class, Tsinghua University. Before college, I competed in informatics olympiads.

Education

New York University Courant Institute, Sept. 2023 – Present
Ph.D. in Computer Science, advised by Prof. Jinyang Li and Prof. Aurojit Panda
Tsinghua University, Sept. 2019 – June 2023
B.Eng. in Computer Science (Honored Yao Class)

Industry Internships

NVIDIA Spatial Intelligence Lab, May 2025 – Dec. 2025
Research Scientist Intern
Project: Large-scale 3D point cloud perception system built on Point Transformer V3 and fVDB.
Microsoft DeepSpeed, May 2024 – Aug. 2024
Research Scientist Intern
Project: Fine-grained in-kernel overlapping of GEMM and collective communication.

Publications

Scaling Point-based Differentiable Rendering for Large-scale Reconstruction
Hexu Zhao, Xiaoteng Liu, Xiwen Min, Jianhao Huang, Youming Deng, Yanfei Li, Ang Li, Jinyang Li, Aurojit Panda
[arXiv] (The PhD work I’m proudest of intellectually, despite rejections from SOSP, OSDI, and ASPLOS.)
CLM: Removing the GPU Memory Barrier for 3D Gaussian Splatting
*Hexu Zhao, *Xiwen Min, Xiaoteng Liu, Moonjun Gong, Yiming Li, Ang Li, Saining Xie, Jinyang Li, Aurojit Panda
ASPLOS 2026 [arXiv] [Code] [Project]
On Scaling Up 3D Gaussian Splatting Training
Hexu Zhao, *Haoyang Weng, *Daohan Lu, Ang Li, Jinyang Li, Aurojit Panda, Saining Xie
ICLR 2025 Oral [arXiv] [Code] [Project]
On Optimizing the Communication of Model Parallelism
*Yonghao Zhuang, *Hexu Zhao, Lianmin Zheng, Zhuohan Li, Eric P. Xing, Qirong Ho, Joseph E. Gonzalez, Ion Stoica, Hao Zhang
MLSys 2023 [arXiv]
WildCity: A Real-World City-Scale Testbed for Rendering, Simulation, and Spatial Intelligence
Xiangyu Han, Mengyu Yang, Jiaqi Li, Bowen Chang, Ziyu Chen, Hexu Zhao, Rahul Kumar Agrawal, Anthony Rodriguez, Rajani Acharya, Fiona Hua, Marco Pavone, Chen Feng, Yiming Li
ECCV 2026 [Paper] [Project]
Fully Hyperbolic Neural Networks
Weize Chen, Xu Han, Yankai Lin, Hexu Zhao, Zhiyuan Liu, Peng Li, Maosong Sun, Jie Zhou
ACL 2022 [arXiv]
Explainable Deep Learning Framework Incorporating Medical Knowledge for Insulin Titration in Diabetes
*Haowei He, *Zhen Ying, Biao Li, Yujuan Fan, Ping Wang, Jiaping Lu, Liming Wu, Hexu Zhao, Yanying Guo, Guangyu Wang, Yang Yuan, Ying Chen, Xiaoying Li
Communications Medicine 2026 [Paper]

My Past

I was born in a small city (Zibo, Shandong) in China.

In my childhood, I only cared about playing, carefree, with more than enough joy and love from my parents.

In middle school, I had exactly the same joys and troubles as any other Chinese teenager going through puberty.

In high school, I set out to become Zibo's first-ever National Olympiad in Informatics gold medalist — by self-studying and giving up on Gaokao prep entirely — and after going through enormous pain and pressure, I made it. I got to experience being a labeled "local legend," which pretty much cured me of caring about external titles and rankings.

In college, with the resources of Tsinghua Yao Class, and driven by broad curiosity, I tried everything I had the chance to try. But I never went deep on anything. Just enough to get by (for example, I graduated). I also met all kinds of people and had many unforgettable conversations. But to be honest, it was a time of feeling a bit lost.

During my PhD at NYU, I experienced another stretch of passion. I found a niche in Gaussian Splatting systems that most people had overlooked, and went all in for 2.5 years. I got to experience the full cycle of research and the AI academic world, pretty interesting. I also had a great life in NYC along the way.

Each of these shaped me in ways I didn't expect. I can't wait to see what finds me next...