PeopleofSuperintelligence

Researcher

Kaiwen Jiang

Also known as Kevin

Kaiwen Jiang is a computer-graphics researcher whose work spans neural rendering and 3D face synthesis. He earned a bachelor's degree at Beijing Jiaotong University. His 2022 NeRFFaceEditing paper separated face geometry and appearance; his 2023 NeRFFaceLighting paper modeled portrait lighting with neural radiance fields. Later work applied Gaussian splatting to sparse-view synthesis and surface reconstruction, then animated expressive 3D head avatars at over 100 frames per second. He received UC San Diego's MS Research Award in 2025 and, as of October 2026, is a computer science PhD candidate there.

Works
7
Citations
8
h-index
2

Publication numbers from OpenAlex, as of 15 October 2026.

Timeline

  1. 2022

    NeRFFaceEditing for face control

    His bachelor's degree is from Beijing Jiaotong University. In 2022, he coauthors NeRFFaceEditing, which lets users edit facial geometry and appearance independently in a pretrained tri-plane-based neural radiance field.

  2. 2023

    NeRFFaceLighting for portrait relighting

    With Shu-Yu Chen, Hongbo Fu and Lin Gao, he publishes NeRFFaceLighting in 2023. The method learns an implicit representation for portrait lighting from a pretrained tri-plane model.

  3. 2024

    Sparse-view synthesis with Gaussian splats

    In 2024, he coauthors A Construct-Optimize Approach to Sparse View Synthesis without Camera Pose. The method combines 3D Gaussian splatting, monocular depth and image correspondences to reconstruct scenes with missing camera poses.

  4. 2025

    Reconstructing surfaces with Gaussian surfels

    At CVPR 2025, he coauthors Geometry Field Splatting with Gaussian Surfels. The method uses Gaussian surfels to render geometry fields and reconstruct opaque surfaces.

  5. June 2025

    A computer science research award

    UC San Diego awards him the Computer Science Department's MS Research Award in June 2025. He is due back as a computer science PhD student under Ravi Ramamoorthi.

  6. 2026

    Expressive avatars over 100 FPS

    At ECCV 2026, he coauthors Instant Expressive Gaussian Head Avatars at Over 100 FPS. Its feed-forward encoder turns a single in-the-wild image into an animatable 3D representation, with animation and pose control running at 107.31 FPS.

  7. As of October 2026

    PhD candidate at UC San Diego

    As of October 2026, he is a computer science PhD candidate at UC San Diego. His research interests span computer graphics and virtual reality.

Sources and credits

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

  1. Kaiwen(Kevin)'s Homepage raymondjiangkw.github.io. Checked 15 October 2026.
  2. NeRFFaceEditing: Disentangled Face Editing in Neural Radiance Fields arxiv.org. Checked 15 October 2026.
  3. NeRFFaceLighting: Implicit and Disentangled Face Lighting Representation scholars.cityu.edu.hk. Checked 15 October 2026.
  4. A Construct-Optimize Approach to Sparse View Synthesis without Camera Pose arxiv.org. Checked 15 October 2026.
  5. Geometry Field Splatting with Gaussian Surfels openaccess.thecvf.com. Checked 15 October 2026.
  6. Congratulations to Kaiwen Jiang jacobsschool.ucsd.edu. Checked 15 October 2026.
  7. Instant Expressive Gaussian Head Avatars at Over 100 FPS research.nvidia.com. Checked 15 October 2026.