PeopleofSuperintelligence

Researcher

Hanzi Mao

Hanzi Mao is a research scientist whose computer-vision work runs from ConvNeXt and Segment Anything to world models for Physical AI. They studied at Huazhong University of Science and Technology and did doctoral work at Texas A&M University, then joined Facebook AI Research in 2021. In 2022 they coauthored ConvNeXt and ViTDet, and in 2023 Segment Anything, which received an ICCV 2023 Best Paper Honorable Mention. From 2023 to 2025 they were a research scientist at Nvidia, serving as research tech lead on Cosmos world foundation models. They are a research scientist on the Google DeepMind Robotics team.

Badges

Timeline

  1. 2009–2015

    Studies at Huazhong University

    From 2009 to 2013, they are an undergraduate at Huazhong University of Science and Technology, and from 2013 to 2015 a master's student there.

  2. 2019–2020

    Machine learning for soil moisture

    A PhD student at Texas A&M University, they publish a gap-filled soil moisture product for SMAP/Sentinel-1 in Water Resources Research in 2019, with Dhruva Kathuria, Nick Duffield and Binayak Mohanty. In 2020, an ICDM paper with Mao as first author, written with Xi Liu, Nick Duffield, Hao Yuan, Shuiwang Ji and Binayak Mohanty, proposes a semi-supervised, attention-based deep representation model for geo-spatiotemporal data.

  3. 2021–2023

    Research scientist at FAIR

    From 2021 to 2023, they are a research scientist at Facebook AI Research (FAIR).

  4. 2022

    ConvNeXt and ViTDet

    In 2022, they coauthor "A ConvNet for the 2020s" with Zhuang Liu, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell and Saining Xie, which introduces the ConvNeXt family of pure ConvNet models and reaches 87.8% ImageNet top-1 accuracy. The plain-Vision-Transformer detector ViTDet, written with Yanghao Li, Ross Girshick and Kaiming He, reaches up to 61.3 AP_box on COCO.

  5. 2023

    Segment Anything at ICCV

    In April 2023, they are one of twelve authors of Segment Anything, which introduces the promptable Segment Anything Model and the SA-1B dataset of over 1 billion masks on 11M images. The paper receives an ICCV 2023 Best Paper Honorable Mention.

  6. 2023–2025

    Research scientist at Nvidia

    From 2023 to 2025, they are a research scientist at Nvidia Deep Imagination Research, building world models for Physical AI.

  7. 2025

    Cosmos world foundation models

    In January 2025, the Cosmos World Foundation Model Platform for Physical AI appears on arXiv with Mao among its authors; the platform covers a video curation pipeline, pre-trained world foundation models and video tokenizers, and is open-sourced. In March 2025, Cosmos-Transfer1, a conditional world generation model that takes spatial control inputs such as segmentation, depth and edge, appears with Mao among its authors. Their role on the Cosmos platform, Cosmos-Transfer1 and Cosmos-Predict2 is research tech lead.

  8. As of October 2026

    Research scientist at DeepMind

    They are a research scientist on the Google DeepMind Robotics team. Their aim is to build intelligent machines that perceive, reason and act in the real world, and they believe foundation models trained on large-scale, diverse data can speed progress toward that goal.

Career

  1. NowGoogle DeepMindResearch Scientist
  2. 2023–2025NVIDIAResearch Scientist; Research Tech Lead
  3. 2021–2023FAIR (Meta)Research Scientist

Education

  1. 2020Texas A&M UniversityPhD
  2. 2015Huazhong University of Science and TechnologyMaster's
  3. 2013Huazhong University of Science and TechnologyUndergraduate

From the biography.

  1. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Hanzi Mao. 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.
  2. 16% of AI career moves in the 2020s went from a university to a companyIncludes Hanzi Mao. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  3. 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.
  4. 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.

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 Hanzi Mao yet. Send us a free-licensed photo.

  1. Hanzi Mao (Google): Physical AI, World models — Lacuna lacuna.tiptreesystems.com. Checked 21 October 2026.
  2. Hanzi Mao hanzimao.me. Checked 21 October 2026.
  3. [2201.03545] A ConvNet for the 2020s arxiv.org. Checked 21 October 2026.
  4. [2203.16527] Exploring Plain Vision Transformer Backbones for Object Detection arxiv.org. Checked 21 October 2026.
  5. [2304.02643] Segment Anything arxiv.org. Checked 21 October 2026.
  6. [2501.03575] Cosmos World Foundation Model Platform for Physical AI arxiv.org. Checked 21 October 2026.
  7. [2503.14492] Cosmos-Transfer1: Conditional World Generation with Adaptive Multimodal Control arxiv.org. Checked 21 October 2026.
  8. Hanzi Mao - Google Scholar scholar.google.com. Checked 21 October 2026.