Researcher

Tianyi Xie

Tianyi Xie is a computer graphics and embodied AI researcher who uses physical simulation with generative models to create motion, garments and animation. They worked at ByteDance AI Lab and Netflix Research before joining NVIDIA Research, and coauthored PhysGaussian, Dress-1-to-3 and PhysAnimator. At UCLA, they defended a computer science Ph.D. on physics-informed generative modeling and received the Computer Science department's Outstanding Graduate Student Research Award in 2026. Since April 2026, they have been a Research Scientist in NVIDIA Research's Data-Driven AI for Robotics team, where they coauthored GRAIL for humanoid loco-manipulation.

Now at
NVIDIA
Works
10
Citations
381
h-index
5
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2020–2021

    ByteDance AI Lab internship

    ByteDance AI Lab in Shanghai employs Xie as a research intern from April 2020 to July 2021.

  2. June 2024–November 2024

    Netflix Research internship

    Netflix Research takes Xie on as a research intern in Los Gatos in June 2024; the role runs through November 2024.

  3. 2024

    PhysGaussian at CVPR

    PhysGaussian is a CVPR 2024 paper coauthored by Xie; it integrates Newtonian dynamics into 3D Gaussian kernels to synthesize novel motion.

  4. February 2025

    The Dress-1-to-3 pipeline

    The Dress-1-to-3 pipeline, coauthored by Xie in February 2025, reconstructs physics-plausible, simulation-ready garments from an in-the-wild image.

  5. 2025

    PhysAnimator at CVPR

    PhysAnimator combines physics-based simulations with generative models to animate static anime illustrations; Xie coauthors the CVPR 2025 paper.

  6. June 2025–April 2026

    NVIDIA Research internship

    NVIDIA Research employs Xie as a research intern in Santa Clara from June 2025 to April 2026.

  7. February 2026–April 2026

    A UCLA PhD and research award

    A February 2026 UCLA computer science Ph.D. defense centers on Xie’s thesis, “Physics-Informed Methods for Generative Modeling and Dynamics.” They hold an M.S. in computer science from UCLA and a B.Eng. in software engineering from Shanghai Jiao Tong University; Chenfanfu Jiang and Demetri Terzopoulos advise the doctoral research. UCLA CS awards Xie the Outstanding Graduate Student Research Award in April 2026.

  8. Since April 2026

    Research Scientist at NVIDIA

    In April 2026, Xie joins NVIDIA Research as a Research Scientist on its Data-Driven AI for Robotics (DAIR) team. Their work connects physics-based simulation and generative AI across computer graphics, embodied AI and robotics. In June 2026, Xie coauthors GRAIL, a framework for generating humanoid loco-manipulation from 3D assets and video priors.

Career

  1. Since 2026NVIDIAResearch Intern, Research Scientist
  2. 2024NetflixResearch Intern
  3. 2020–2021ByteDanceResearch Intern

Education

  1. 2026University of California, Los AngelesPhD
  2. Shanghai Jiao Tong UniversityB.Eng.

Sources and credits

Photo: We have no published photo of Tianyi Xie yet. Send us a free-licensed photo.

  1. Tianyi Xie | NVIDIA Research xpandora.github.io. Checked 22 October 2026.
  2. CVPR 2024 Open Access Repository openaccess.thecvf.com. Checked 22 October 2026.
  3. [2502.03449] Dress-1-to-3: Single Image to Simulation-Ready 3D Outfit with Diffusion Prior and Differentiable Physics arxiv.org. Checked 22 October 2026.
  4. CVPR 2025 Open Access Repository openaccess.thecvf.com. Checked 22 October 2026.
  5. UCLA Artificial Intelligence & Visual Computing Lab math.ucla.edu. Checked 22 October 2026.
  6. Graduate Students and Postdocs of Professor Demetri Terzopoulos web.cs.ucla.edu. Checked 22 October 2026.
  7. Tianyi Xie | Research research.nvidia.com. Checked 22 October 2026.
  8. DAIR @ NVIDIA Research research.nvidia.com. Checked 22 October 2026.
  9. GRAIL: Generating Humanoid Loco-Manipulation from 3D Assets and Video Priors | DAIR @ NVIDIA Research research.nvidia.com. Checked 22 October 2026.