PeopleofSuperintelligence

Researcher

José Lezama

Also known as Jose Lezama

José Lezama is a computer-vision researcher whose work spans image representation learning and video generation. He earned a BSc from Universidad de la República and MSc and PhD degrees at École Normale Supérieure Paris-Saclay, and coauthored papers on video segmentation and cross-spectrum face recognition. His research includes Jacobian supervision, Token-Critic and MAGVIT; VideoPoet won ICML's 2024 Best Paper Award. As of October 2026, he is a Staff Research Scientist at Google DeepMind.

Timeline

  1. 2011

    Degrees and video segmentation

    His education includes a BSc from Universidad de la República and MSc and PhD degrees from École Normale Supérieure Paris-Saclay, with Jean-Michel Morel and Gregory Randall as advisors. In 2011, he coauthors “Track to the Future,” a CVPR paper on spatiotemporal video segmentation with long-range motion cues.

  2. 2016

    Cross-spectrum face recognition

    In 2016, he coauthors “Not Afraid of the Dark,” a near-infrared and visible-light face-recognition paper with Qiang Qiu and Guillermo Sapiro. It combines cross-spectral hallucination with low-rank embedding.

  3. 2019

    Jacobian supervision for images

    In 2019, he introduces Jacobian supervision, which grows a latent code while constraining how image output changes with disentangled variables.

  4. Before 2021

    Assistant professor in Uruguay

    Before 2021, he works as an Assistant Professor at Universidad de la República in Uruguay.

  5. 2022

    Token-Critic guides sampling

    In 2022, he coauthors Token-Critic, an auxiliary model that guides non-autoregressive transformer sampling by selecting which image tokens to accept or reject and resample.

  6. 2023

    MAGVIT and discrete diffusion

    In 2023, he coauthors MAGVIT, a masked generative video transformer for multi-task video generation. He also coauthors a predictor-corrector method for discrete diffusion that improves class-conditional image generation on ImageNet.

  7. 2024

    VideoPoet and W.A.L.T. models

    In 2024, he coauthors VideoPoet, a multimodal language model for zero-shot video generation, and W.A.L.T., a diffusion transformer for photorealistic video generation. ICML gives VideoPoet its Best Paper Award. He is a Senior Research Scientist at Google DeepMind.

  8. As of October 2026

    Google DeepMind research

    As of October 2026, he is a Staff Research Scientist at Google DeepMind, working on VideoPoet, W.A.L.T., Veo 2 and Veo 3.

Career

  1. Since 2024Google DeepMindSenior Research Scientist
  2. University of the RepublicAssistant Professor

Education

  1. University of the RepublicBSc
  2. École Normale Supérieure Paris-SaclayMSc

From the biography.

  1. Google had the most people shaping AI every year from 2014 to 2025Includes Google. 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.
  2. Google has the highest combined h-index among AI organizationsIncludes Google. 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.
  3. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. 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.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Google. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

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 José Lezama yet. Send us a free-licensed photo.

  1. José Lezama - Staff Research Scientist jlezama.github.io. Checked 21 October 2026.
  2. Track to the Future : Spatio-temporal Video Segmentation di.ens.fr. Checked 21 October 2026.
  3. Not Afraid of the Dark: NIR-VIS Face Recognition via Cross-spectral Hallucination and Low-rank Embedding arxiv.org. Checked 21 October 2026.
  4. Overcoming the Disentanglement vs Reconstruction Trade-Off via Jacobian Supervision mlanthology.org. Checked 21 October 2026.
  5. Plenary Talks – IBERAMIA’2024 iberamia.org. Checked 21 October 2026.
  6. Improved Masked Image Generation with Token-Critic research.google. Checked 21 October 2026.
  7. MAGVIT: Masked Generative Video Transformer research.google. Checked 21 October 2026.
  8. Predictor-Corrector Sampling for Discrete Diffusion Models research.google. Checked 21 October 2026.
  9. VideoPoet: A Large Language Model for Zero-Shot Video Generation research.google. Checked 21 October 2026.
  10. Photorealistic Video Generation with Diffusion Models research.google. Checked 21 October 2026.
  11. ICML 2024 Awards icml.cc. Checked 21 October 2026.