Shuhan Tan

Hi there! I am a Research Scientist in the Autonomous Systems and Physical AI Research (ASPIRE) Group at NVIDIA Research. I received my PhD in Computer Science from The University of Texas at Austin in March 2026, advised by Philipp Krähenbühl.

My research focuses on building physical AI agents for robotics and self-driving.

Email  /  CV  /  Scholar  /  LinkedIn

PontTuset
Education

The University of Texas at Austin
PhD in Computer Science • Aug. 2021 - Mar. 2026

Sun Yat-Sen University
B.E. in Computer Science • Sep. 2016 - Jun. 2021
Ranking: 2/189

Research Experience

NVIDIA Research
Autonomous Systems and Physical AI Research (ASPIRE) Group
Research Scientist • Jun. 2026 - Present
Working on physical AI agents for robotics and self-driving.

Waymo Research
Research Intern • Jun. 2024 - Nov. 2024
With Chiyu "Max" Jiang

NVIDIA Research
Autonomous Systems and Physical AI Research (ASPIRE) Group
Research Intern • Aug. 2023 - Jun. 2024;
Research Intern • Jun. 2025 - May 2026
With Boris Ivanovic , Xinshuo Weng , Yuxiao Chen ,
            Boyi Li , Yulong Cao , Marco Pavone

The University of Texas at Austin
Research Assistant • Jan. 2023 - Mar. 2026
PhD Advisor: Philipp Krähenbühl
Research Assistant • Aug. 2021 - Dec. 2022
With Kristen Grauman

The Chinese University of Hong Kong
Research Assistant • Sep. 2020 - Mar. 2021
With Bolei Zhou

Uber ATG Toronto
Research Intern • Sep. 2019 - Aug. 2020
With Raquel Urtasun, Shenlong Wang

Boston University
Research Assistant • July 2019 - Sep. 2019
With Kate Saenko, Xingchao Peng

Sun Yat-Sen University
Undergraduate Researcher • Sep. 2017 - June 2019
With Wei-Shi Zheng

News

[06/2026] I joined the Autonomous Systems and Physical AI Research (ASPIRE) Group at NVIDIA Research as a Research Scientist!

[03/2026] I graduated from The University of Texas at Austin with a PhD in Computer Science!

[02/2026] LCDrive got accepted to CVPR 2026! Check it out!

[10/2025] I was selected as an Outstanding Reviewer of ICCV 2025 (Top 3%)!

[06/2025] InfGen got accepted to ICCV 2025! Check it out!

[05/2025] RIPT-VLA released! Check it out!

[03/2025] SceneDiffuser++ got accepted to CVPR 2025!

Talks
PontTuset Towards realistic, controllable and reactive
traffic simulation



Long-term Human Motion Prediction Workshop
ICRA 2024. Yokohama, Japan.


Slides / Workshop

Publications
LCDrive Latent Chain-of-Thought World Modeling for End-to-End Driving.

Shuhan Tan, Kashyap Chitta, Yuxiao Chen, Ran Tian, Yurong You, Yan Wang, Wenjie Luo, Yulong Cao, Philipp Krähenbühl, Marco Pavone, Boris Ivanovic

CVPR 2026.

Paper

A Vision-Language-Action model for autonomous driving that uses latent chain-of-thought reasoning with a learned world model for improved trajectory planning.

PontTuset Long-term Traffic Simulation with Interleaved Autoregressive Motion and Scenario Generation.


Xiuyu Yang*, Shuhan Tan*, Philipp Krähenbühl

ICCV 2025.

Page / Paper / Code

A unified next-token prediction model that performs interleaved closed-loop motion simulation and scene generation for stable long-term traffic simulation.

PontTuset RIPT-VLA: Interactive Post-Training for Vision-Language-Action Models.

Shuhan Tan, Kairan Dou, Yue Zhao, Philipp Krähenbühl

arxiv 2505.17016 preprint.

Paper / Page / Code / Model

Interactive post-training for any pretrained Vision-Language-Action (VLA) model using only sparse binary success rewards.

PontTuset SceneDiffuser++: City-Scale Traffic Simulation via a Generative World Model.

Shuhan Tan, John Lambert, Hong Jeon, Sakshum Kulshrestha, Yijing Bai, Jing Luo, Dragomir Anguelov, Mingxing Tan, Chiyu Max Jiang

CVPR 2025.

Paper / Page

Diffusion-based World Model for city-scale traffic simulation.

PontTuset Promptable Closed-loop Traffic Simulation.


Shuhan Tan, Boris Ivanovic, Yuxiao Chen, Boyi Li, Xinshuo Weng, Yulong Cao, Philipp Krähenbühl, Marco Pavone


CoRL 2024.

Paper / Project page / Code / Video / Colab Demo / Dataset

Simulate multi-agent interactions with multimodal prompts.

PontTuset Language Conditioned Traffic Generation.


Shuhan Tan, Boris Ivanovic, Xinshuo Weng, Marco Pavone, Philipp Krähenbühl


CoRL 2023.

Paper / Project page / Code / Video / Colab Demo

Traffic scene generation with language condition using LLM.

PontTuset EgoDistill: Egocentric Head Motion Distillation for Efficient Video Understanding.

Shuhan Tan, Tushar Nagarajan, Kristen Grauman

NeurIPS 2023.

Paper / Project page

Efficient egocentric video understanding with head motion data from IMU.

PontTuset TrafficGen: Learning to Generate Diverse and Realistic Traffic Scenarios.

Lan Feng, Quanyi Li, Zhenghao Peng, Shuhan Tan, Bolei Zhou

ICRA 2023.

Paper / Project page/ Video

Synthesis new traffic scenario and replay in simulation.

PontTuset Improving the Fairness of Deep Generative Models without Retraining.

Shuhan Tan, Yujun Shen, Bolei Zhou.

arXiv.2012.04842 preprint

Paper / Project page / Colab

Mitigate biases of GAN models without retraining.

PontTuset SceneGen: Learning to Generate Realistic Traffic Scenes.

Shuhan Tan*, Kelvin Wong*, Shenlong Wang, Sivabalan Manivasagam, Mengye Ren, Raquel Urtasun.

CVPR 2021.

Paper / Video / Bibtex

Generate realistic traffic scences automatically.

PontTuset Class-imbalanced Domain Adaptation: An Empirical Odyssey.

Shuhan Tan, Xingchao Peng, Kate Saenko.


TASK-CV Workshop, ECCV 2020.

Paper / Bibtex

Align feature distributions across domains while the label distributions of the two domains are also different.

PontTuset LidarSIM: Realistic LiDAR Simulation by Leveraging the Real World

Sivabalan Manivasagam, Shenlong Wang, Kelvin Wong, Wenyuan Zeng, Mikita Sazanovich, Shuhan Tan, Bin Yang, Wei-Chiu Ma and Raquel Urtasun.

CVPR 2020 (Oral).

Paper / Supplement / Bibtex

Realistic sensor simulation for LiDAR.

PontTuset

Biomarker Localization by Combining CNN Classifier and Generative Adversarial Network

Rong Zhang, Shuhan Tan, Ruixuan Wang, Siyamalan Manivannan, Jingjing Chen, Haotian Lin, Wei-Shi Zheng.

MICCAI 2019.

Paper / Bibtex

We proposed a novel deep neural network architecture to effectively localize potential biomarkers in medical images, when only the image-level labels are available during model training.

PontTuset

Weakly Supervised Open-set Domain Adaptation by Dual-domain Collaboration

Shuhan Tan, Jiening Jiao, Wei-Shi Zheng

CVPR 2019.

Paper / Supplement / Bibtex

We proposed a practical weakly supervised setting for open-set domain adaptation, where two scarcely-labeled domains collaboratively learn from each other.

Invited for presentation at WebVision 2019.

Selected Honors

Distinguished Graduate Thesis, Sun Yat-sen University

Patent

Systems and Methods for Simulating Traffic Scenes. US Patent App. 17/528,277.

Shuhan Tan, Kelvin Ka Wing Wong, Shenlong Wang, Sivabalan Manivasagam, Mengye Ren, Raquel Urtasun
Acedemic Service

Reviewer: CVPR, ECCV, ICCV, ICLR, ICML, ICRA, IJCV, NeruIPS, RA-L, TCSVT, TMI.
Cat Service
PontTuset Shufa, aka Maomi
British Longhair Boy😺



Born in 2022, California.



Instagram / Wiki

My little colleague, who always sleeps on my desk,
put his paw on the F5 key to prevent vscode debugging.


Updated Sep 2026.

Special thanks to Jon Barron for website template.