Researcher

Justin Johnson

Also known as Justin M. Johnson

Justin Johnson is a computer vision researcher whose work connects images, language and 3D structure, and a co-founder of the spatial-intelligence company World Labs. He earned a Stanford PhD under Fei-Fei Li in 2018, after submitting the perceptual-loss paper for real-time style transfer and the CLEVR visual-reasoning dataset to arXiv in 2016. By 2019, he was an assistant professor at the University of Michigan following a research scientist post at Facebook AI Research, and he coauthored Mesh R-CNN. In 2024, he co-founded World Labs with Li, Ben Mildenhall and Christoph Lassner, and the company launched its Marble world model in November 2025. As of October 2026, he is a World Labs co-founder; the company signed an agreement in September 2026 to join AMD, and he and Mildenhall are to continue to lead its team with Li.

Timeline

  1. 2016

    Perceptual losses and CLEVR

    In March 2016, he submits a paper on perceptual losses for real-time style transfer and super-resolution to arXiv with Alexandre Alahi and Li Fei-Fei; the feed-forward network gives results similar to the optimization-based style-transfer method and runs three orders of magnitude faster. In December 2016, he submits CLEVR with Bharath Hariharan, Laurens van der Maaten, Li Fei-Fei, C. Lawrence Zitnick and Ross Girshick, a diagnostic dataset that tests visual reasoning abilities.

  2. 2017

    Co-teaching CS231n at Stanford

    He is one of three instructors, with Fei-Fei Li and Serena Yeung, of Stanford's CS231n course on convolutional neural networks for visual recognition in spring 2017.

  3. 2018

    Stanford PhD under Fei-Fei Li

    He completes a Stanford computer science PhD in 2018 under Fei-Fei Li, with the dissertation Compositional Visual Intelligence. It covers dense captions and paragraphs for images, a visual question answering benchmark with a compositional model, and image generation from scene graphs; Noah Goodman and Christopher Ré serve on his committee.

  4. 2019

    Michigan faculty and Mesh R-CNN

    By October 2019, he is an assistant professor of computer science and engineering at the University of Michigan in Ann Arbor, after working as a research scientist at Facebook AI Research. In 2019, he coauthors Mesh R-CNN with Georgia Gkioxari and Jitendra Malik, a system that detects objects in real-world images and produces a triangle mesh for the 3D shape of each one.

  5. 2022

    A MICDE Catalyst Grant

    In 2022, as an assistant professor of electrical engineering and computer science, he receives a MICDE Catalyst Grant at Michigan. He and Ada Eban-Rothschild, an assistant professor of psychology, plan a computer vision toolbox that automatically tracks multiple animals and classifies their behaviors in complex social environments.

  6. 2024

    Co-founding World Labs

    With Fei-Fei Li, Ben Mildenhall and Christoph Lassner, he co-founds World Labs in 2024, a startup that builds AI with spatial intelligence, the ability to reason about and act within three-dimensional environments.

  7. November 2025

    Marble, a 3D world model

    In November 2025, World Labs launches Marble, its first commercial world model, which turns text prompts, photos, videos, 3D layouts or panoramas into editable, downloadable 3D environments. He and Li discuss the launch on the Latent Space podcast in November 2025.

  8. As of October 2026

    World Labs joins AMD

    In September 2026, World Labs signs a definitive agreement to join AMD in an all-stock deal valued at about $8.2 billion, expected to close by the end of 2026. He and Ben Mildenhall are to work with Li to continue to lead the World Labs team as it joins AMD. He is also an adjunct professor at the University of Michigan.

Career

  1. Since 2024World LabsCo-founder
  2. Since 2019University of MichiganAssistant Professor
  3. 2017Stanford UniversityInstructor
  4. FAIR (Meta)Research Scientist

Education

  1. 2018Stanford UniversityPhD

Sources and credits

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

  1. Perceptual Losses for Real-Time Style Transfer and Super-Resolution arxiv.org. Checked 21 October 2026.
  2. CLEVR: A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoning arxiv.org. Checked 21 October 2026.
  3. Stanford University CS231n: Convolutional Neural Networks for Visual Recognition cs231n.stanford.edu. Checked 21 October 2026.
  4. Compositional visual intelligence | Stanford Digital Repository purl.stanford.edu. Checked 21 October 2026.
  5. From Recognition to Reasoning (Seoul National University Computer Science and Engineering seminar) cse.snu.ac.kr. Checked 21 October 2026.
  6. Mesh R-CNN arxiv.org. Checked 21 October 2026.
  7. Catalyst Grants enable computational-science innovations | U-M LSA Department of Psychology prod.lsa.umich.edu. Checked 21 October 2026.
  8. Fei-Fei Li - Wikipedia en.wikipedia.org. Checked 21 October 2026.
  9. About | World Labs worldlabs.ai. Checked 21 October 2026.
  10. AMD agrees to buy Fei-Fei Li's World Labs for $8.2bn in stock thenextweb.com. Checked 21 October 2026.
  11. Fei-Fei Li's World Labs speeds up the world model race with Marble, its first commercial product | TechCrunch techcrunch.com. Checked 21 October 2026.
  12. After LLMs: Spatial Intelligence and World Models — Fei-Fei Li & Justin Johnson, World Labs latent.space. Checked 21 October 2026.
  13. World Labs is Joining AMD | World Labs worldlabs.ai. Checked 21 October 2026.
  14. Justin Johnson | TWIML - The Voice of Machine Learning & AI twimlai.com. Checked 21 October 2026.