PeopleofSuperintelligence

Researcher

Kaixuan Wei

Kaixuan Wei is a computer scientist who builds learning-based methods for computational imaging, photography and optics. He earned a B.S. in electronic engineering in 2018 and an M.S. in computer science in 2021 at Beijing Institute of Technology, with research visits to Microsoft Research Asia in 2018 and the University of Cambridge in 2019. His tuning-free plug-and-play proximal algorithm for inverse imaging problems won an ICML Outstanding Paper award in 2020 and appeared in the Journal of Machine Learning Research in 2022, followed by the ∇-Prox differentiable optimization compiler at SIGGRAPH 2023. In 2023, he posted Spatially Varying Nanophotonic Neural Networks, published in Science Advances in 2024, which embedded optical computation into flat camera optics. Since January 2024, he is a PhD student at King Abdullah University of Science and Technology under Wolfgang Heidrich in the KAUST Visual Computing Center.

Badges

Timeline

  1. 2014–2018

    Electronics degree at BIT

    He studies electronic engineering at Beijing Institute of Technology from September 2014 to June 2018 and receives his B.S. degree in 2018.

  2. 2018–2021

    Master's degree and research visits

    From September 2018 to June 2021, he studies computer science at Beijing Institute of Technology and completes an M.S. degree in 2021. He interns at Microsoft Research Asia from April 2018 to November 2018 and visits the Cambridge Image Analysis group at the University of Cambridge from July 2019 to September 2019.

  3. 2020

    Low-light noise modeling at CVPR

    In 2020, he is first author of A Physics-based Noise Formation Model for Extreme Low-light Raw Denoising, accepted as an oral presentation at CVPR 2020. Coauthored with Ying Fu, Jiaolong Yang and Hua Huang, the paper models the characteristics of CMOS photosensors to synthesize realistic training samples.

  4. 2020

    An ICML Outstanding Paper award

    At ICML 2020, he shares an Outstanding Paper award with Angelica Aviles-Rivero, Jingwei Liang, Ying Fu, Carola-Bibiane Schönlieb and Hua Huang for Tuning-free Plug-and-Play Proximal Algorithm for Inverse Imaging Problems. Their method trains a policy network with model-free and model-based deep reinforcement learning to set the penalty parameter, denoising strength and terminal time automatically.

  5. 2022

    TFPnP in JMLR

    The journal version, TFPnP: Tuning-free Plug-and-Play Proximal Algorithms with Applications to Inverse Imaging Problems, appears in the Journal of Machine Learning Research in 2022. It demonstrates results on compressed sensing MRI, sparse-view CT, single-photon imaging and phase retrieval.

  6. 2023

    Differentiable proximal solvers at SIGGRAPH

    In 2023, he coauthors ∇-Prox, a domain-specific modeling language and compiler presented at SIGGRAPH 2023 that transforms optimization problems into differentiable proximal solvers.

  7. 2023–2024

    Nanophotonic neural networks in camera optics

    In August 2023, he posts Spatially Varying Nanophotonic Neural Networks with Xiao Li, Johannes Froech, Praneeth Chakravarthula, James Whitehead, Ethan Tseng, Arka Majumdar and Felix Heide, and the paper appears in Science Advances in 2024. The system embeds parallelized optical computation into flat camera optics during capture and reaches 72.76% accuracy on CIFAR-10, ahead of AlexNet's 72.64%.

  8. Since 2024

    PhD student at KAUST

    He begins a PhD in computer science at King Abdullah University of Science and Technology in January 2024, working under Wolfgang Heidrich in the KAUST Visual Computing Center on computational imaging, photography and optics. In 2025, he coauthors Collaborative On-Sensor Array Cameras at SIGGRAPH 2025 and is first author, with Heidrich and Felix Heide among the coauthors, of Large-Area Fabrication-Aware Computational Diffractive Optics at SIGGRAPH Asia 2025.

Career

  1. 2019University of Cambridge
  2. 2018Microsoft Research AsiaIntern

Education

  1. From 2024King Abdullah University of Science and TechnologyPhD student
  2. 2021Beijing Institute of TechnologyM.S.
  3. 2018Beijing Institute of TechnologyB.S.

From the biography.

  1. Among employers with 30+ full-time cases, OpenAI leads primary technical role H-1B pay at $310,000, and Amazon, Google and Microsoft file 57% of all casesIncludes Microsoft. Across 440,669 certified US H-1B filings from 44 employers (FY2020–FY2026), OpenAI leads primary technical role base pay at $310,000 and xAI leads overall base pay at $290,000.
  2. Google had the most people shaping AI every year from 2014 to 2025Includes Microsoft. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes University of Cambridge. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  4. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Microsoft. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.

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 Kaixuan Wei yet. Send us a free-licensed photo.

  1. Kaixuan Wei, PhD Student @ KAUST VCC Imaging (personal homepage) kxwei.net. Checked 22 October 2026.
  2. Kaixuan Wei -- Home Page (KAUST Visual Computing Center) vccimaging.org. Checked 22 October 2026.
  3. People | Cambridge Image Analysis damtp.cam.ac.uk. Checked 22 October 2026.
  4. A Physics-based Noise Formation Model for Extreme Low-light Raw Denoising (arXiv) arxiv.org. Checked 22 October 2026.
  5. ICML 2020 Awards icml.cc. Checked 22 October 2026.
  6. Tuning-free Plug-and-Play Proximal Algorithm for Inverse Imaging Problems (arXiv) arxiv.org. Checked 22 October 2026.
  7. TFPnP: Tuning-free Plug-and-Play Proximal Algorithms with Applications to Inverse Imaging Problems jmlr.org. Checked 22 October 2026.
  8. ∇-Prox: Differentiable Proximal Algorithm Modeling for Large-Scale Optimization – Princeton Computational Imaging Lab light.princeton.edu. Checked 22 October 2026.
  9. Spatially Varying Nanophotonic Neural Networks (arXiv) arxiv.org. Checked 22 October 2026.
  10. Spatially Varying Nanophotonic Neural Networks – Princeton Computational Imaging Lab light.princeton.edu. Checked 22 October 2026.
  11. Collaborative On-Sensor Array Cameras – Princeton Computational Imaging Lab light.princeton.edu. Checked 22 October 2026.
  12. Large-Area Fabrication-Aware Computational Diffractive Optics (arXiv) arxiv.org. Checked 22 October 2026.
  13. Large-Area Fabrication-Aware Computational Diffractive Optics vccimaging.org. Checked 22 October 2026.