PeopleofSuperintelligence

Researcher

Jonathan Lorraine

Jonathan Lorraine is a machine learning researcher at NVIDIA whose work spans nested optimization, text-to-3D generation, video generation and, now, agentic AI for science and engineering. After a double major in computer science and mathematics and an M.Sc.A.C. at the University of Toronto, he published work on hypernetworks in 2018 and on implicit differentiation for hyperparameters in 2020, both with David Duvenaud. At NVIDIA, he was first author of ATT3D at ICCV in 2023, coauthored LATTE3D in 2024, and completed his PhD under Duvenaud in 2024 with a thesis on scalable nested optimization. In 2026, he and his coauthors received an Outstanding Paper Honorable Mention at ICML 2026 for Motion Attribution for Video Generation. He is a senior research scientist in NVIDIA's Learning and Perception Research lab in Toronto.

Now at
NVIDIA
Badges

Timeline

  1. By 2018

    Hypernetworks with Duvenaud

    He double-majors in computer science and mathematics with a minor in economics, working with Dmitry Krass on operations research, then completes an M.Sc.A.C. at the University of Toronto in a joint Statistics and Computer Science program focused on data science. In 2018, he and David Duvenaud post a paper on stochastic hyperparameter optimization through hypernetworks, which trains a neural network to output approximately optimal weights as a function of hyperparameters.

  2. 2020

    Implicit differentiation at AISTATS

    In 2020, he publishes "Optimizing Millions of Hyperparameters by Implicit Differentiation" at AISTATS with Paul Vicol and David Duvenaud. The algorithm combines the implicit function theorem with efficient inverse Hessian approximations to tune millions of hyperparameters by gradient.

  3. 2023

    ATT3D text-to-3D at ICCV

    In 2023, he is first author of ATT3D at ICCV, written with Kevin Xie, Sanja Fidler, James Lucas and others, which trains one model on many text prompts at once so that computation is shared across the prompt set. At NVIDIA's Spatial Intelligence Lab, he works with Fidler on projects from text-to-3D generation to real-time world models for autonomous driving.

  4. March 2024

    The LATTE3D text-to-3D model

    In March 2024, he and Kevin Xie head the author list of LATTE3D, which extends amortized text-to-3D methods such as ATT3D to a significantly larger prompt set. LATTE3D generates 3D objects in 400ms and amortizes both neural field and textured surface generation.

  5. July 2024

    A PhD on nested optimization

    He completes his PhD in machine learning in 2024, advised by David Duvenaud. In July 2024, his thesis "Scalable Nested Optimization for Deep Learning" appears on arXiv, with hyperparameter optimization and generative adversarial networks as motivating examples.

  6. 2026

    An ICML honorable mention

    In 2026, he and his coauthors, including Xindi Wu, Antonio Torralba and Sanja Fidler, receive an Outstanding Paper Honorable Mention at ICML 2026 for Motion Attribution for Video Generation. The paper presents Motive, a gradient-based data attribution framework that identifies which fine-tuning clips improve or degrade temporal dynamics in video generation models.

  7. As of October 2026

    Senior research scientist at NVIDIA

    At NVIDIA's Learning and Perception Research lab in Toronto, he is a senior research scientist working with Jan Kautz on agentic AI for science and engineering.

Career

  1. Since 2023NVIDIASenior research scientist

Education

  1. 2024University of TorontoDouble major in computer science and mathematics

From the biography.

  1. 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.
  2. 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.
  3. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes NVIDIA. 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.
  4. Universities (31%) and startups and other companies (25%) employ the largest shares of 2020–2026 entrantsIncludes NVIDIA. Among 819 people whose first recorded role started in 2020–2026 and who have a known current organization, 60% work in commercial sectors (25% at startups and other companies, 23% at frontier labs and 13% at big tech, with 9 in two), and 31% work at universities, compared with 66% for earlier entrants.

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 Jonathan Lorraine yet. Send us a free-licensed photo.

  1. Jonathan Lorraine | Senior Research Scientist, NVIDIA Research jonlorraine.com. Checked 21 October 2026.
  2. [1802.09419] Stochastic Hyperparameter Optimization through Hypernetworks arxiv.org. Checked 21 October 2026.
  3. Optimizing Millions of Hyperparameters by Implicit Differentiation proceedings.mlr.press. Checked 21 October 2026.
  4. ATT3D: Amortized Text-to-3D Object Synthesis (ICCV 2023 Open Access Repository) openaccess.thecvf.com. Checked 21 October 2026.
  5. [2403.15385] LATTE3D: Large-scale Amortized Text-To-Enhanced3D Synthesis arxiv.org. Checked 21 October 2026.
  6. [2407.01526] Scalable Nested Optimization for Deep Learning arxiv.org. Checked 21 October 2026.
  7. ICML 2026 Awards icml.cc. Checked 21 October 2026.
  8. Motion Attribution for Video Generation (NVIDIA Spatial Intelligence Lab) research.nvidia.com. Checked 21 October 2026.
  9. Jonathan Lorraine | NVIDIA Research research.nvidia.com. Checked 21 October 2026.