PeopleofSuperintelligence

Researcher

Shengze Wang

Shengze Wang is a research scientist at NVIDIA Research who works on 3D vision, world models and telepresence. After studying computer engineering at the University of Illinois at Urbana-Champaign and computer vision at Carnegie Mellon, where he worked on SLAM with Michael Kaess, he completed a PhD on affordable telepresence at UNC Chapel Hill under Henry Fuchs. His papers from those years include EDPLVO, which received the Outstanding Navigation Paper Award at ICRA 2022, Bringing Telepresence to Every Desk in 2023, and the BLADE and Coherent3D papers at CVPR 2025. He defended his thesis in July 2025 and joined NVIDIA Research in September 2025. At NVIDIA, he is first author of Grounded-Exo2Ego, a video diffusion model that generates egocentric video from a single exocentric video and was accepted to NeurIPS 2026.

Now at
NVIDIA
Works
63
Citations
355
h-index
9
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2017–2020

    Degrees at UIUC and CMU

    His research from 2017 to 2020 covers SLAM and classical 3D reconstruction. He studies computer engineering at the University of Illinois at Urbana-Champaign, where he works with Derek Hoiem on 3D reconstruction, and computer vision at Carnegie Mellon University, where he works with Michael Kaess on SLAM.

  2. 2020–2025

    A PhD at UNC Chapel Hill

    He pursues a PhD on affordable telepresence at UNC Chapel Hill, advised by Henry Fuchs. His research in these years spans telepresence, neural rendering, human pose estimation and generative models.

  3. May 2022

    An ICRA navigation paper award

    In May 2022, EDPLVO, a paper on efficient direct point-line visual odometry that he coauthors with Lipu Zhou, Guoquan Huang, Yinian Mao and Michael Kaess, receives the Outstanding Navigation Paper Award at ICRA 2022.

  4. April 2023

    Telepresence for every desk

    In April 2023, he submits Bringing Telepresence to Every Desk to arXiv as first author, with Henry Fuchs among the coauthors. The capture-and-render system needs only 4 consumer-grade RGBD cameras to synthesize free-viewpoint videos of users and their environments.

  5. March 2025

    A VR best paper mention

    In March 2025, he is a coauthor of Multimodal Neural Acoustic Fields for Immersive Mixed Reality, which receives a Best Paper Honorable Mention at IEEE VR 2025. The model learns scene acoustics to render spatial audio from novel viewpoints.

  6. June 2025

    BLADE and Coherent3D at CVPR

    In June 2025, his first-author paper BLADE appears at CVPR 2025. The method recovers a person's body mesh and perspective camera parameters from a single image, including close-range images. A second first-author paper accepted at CVPR 2025, Coherent3D, fuses a canonical 3D prior with per-frame appearance to produce temporally stable 3D portrait video for single-camera telepresence.

  7. July 2025–May 2026

    A UNC doctorate in telepresence

    He defends his PhD thesis at UNC Chapel Hill in July 2025 and is hooded at the commencement in May 2026. His dissertation in the Department of Computer Science, Towards AI-Assisted Cost-Efficient Monitor-Based Telepresence, gathers four papers on view synthesis, 3D video streaming, single-camera portrait reconstruction and single-image human mesh recovery.

  8. Since September 2025

    Research Scientist at NVIDIA

    In September 2025, he joins NVIDIA Research as a Research Scientist in the AI-Mediated Reality and Interaction group led by Shalini De Mello, working on 3D vision, world models, telepresence and physically and socially intelligent robots. In August 2026, he releases Grounded-Exo2Ego on arXiv, a dual-branch video diffusion model that generates egocentric video from a single exocentric video, and in October 2026 it is accepted to NeurIPS 2026.

Career

  1. Since 2025NVIDIAResearch Scientist
  2. 2002–2022Donghua University

Education

  1. 2025University of North Carolina at Chapel HillPhD
  2. University of Illinois Urbana-Champaign
  3. Carnegie Mellon University

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. π-LSAM: LiDAR Smoothing and Mapping With Planes2021 · IEEE International Conference on Robotics and Automation · 57 citations
  2. DPLVO: Direct Point-Line Monocular Visual Odometry2021 · IEEE Robotics and Automation Letters · 44 citations
  3. EDPLVO: Efficient Direct Point-Line Visual Odometry2022 · IEEE International Conference on Robotics and Automation · 43 citations
  4. PLC-LiSLAM: LiDAR SLAM With Planes, Lines, and Cylinders2022 · IEEE Robotics and Automation Letters · 37 citations
  5. Unsupervised Learning of Monocular Depth Estimation with Bundle Adjustment, Super-Resolution and Clip Loss2018 · 19 citations

From OpenAlex, as of 21 October 2026.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Shengze Wang. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  2. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Shengze Wang. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  3. Google has the highest combined h-index among AI organizationsIncludes NVIDIA. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  4. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes NVIDIA. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

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

  1. Shengze Wang | NVIDIA Research research.nvidia.com. Checked 20 October 2026.
  2. Shengze Wang (personal website) mcmvmc.github.io. Checked 20 October 2026.
  3. Bringing Telepresence to Every Desk (arXiv:2304.01197) arxiv.org. Checked 20 October 2026.
  4. BLADE: Single-view Body Mesh Estimation through Accurate Depth Estimation (CVPR 2025 Open Access Repository) openaccess.thecvf.com. Checked 20 October 2026.
  5. Coherent3D: Coherent 3D Portrait Video Reconstruction via Triplane Fusion (arXiv:2412.08684) arxiv.org. Checked 20 October 2026.
  6. Towards AI-Assisted Cost-Efficient Monitor-Based Telepresence (Carolina Digital Repository) doi.org. Checked 20 October 2026.
  7. NVIDIA AI-Mediated Reality and Interaction Research research.nvidia.com. Checked 20 October 2026.
  8. Grounded-Exo2Ego: Structured Semantic Grounding for Robust Exocentric-to-Egocentric Video Generation (arXiv:2608.20534) arxiv.org. Checked 20 October 2026.