PeopleofSuperintelligence

Researcher

Chenhui Zhang

Chenhui Zhang is an AI researcher whose work connects trustworthy language models with machine learning for Earth observation. He earned a computer science B.S. at the University of Illinois Urbana-Champaign before MIT graduate study; he completed the Social and Engineering Systems SM in May 2025 under adviser Sherrie Wang and worked with the Laboratory for Information and Decision Systems and the Earth Intelligence Lab. In 2022, he coauthored a machine-learning study on soil carbon; in 2023, he received a second Fiddler Innovation Fellowship and coauthored a field-level crop residue study. DecodingTrust won a 2023 NeurIPS Outstanding Paper Award, and his 2024 GPT-4V benchmark tested scene understanding, localization, counting and change detection on Earth-observation data. In 2025, he coauthored the MMDT multimodal trust platform and AlphaEarth Foundations for global mapping from sparse labels; as of October 2026, he is a Research Engineer at Google DeepMind working on AI for Science.

Timeline

  1. 2022

    Predicting soil carbon from spectra

    In 2022, he coauthors a Remote Sensing of Environment study that uses soil-library hyperspectral reflectance and machine learning to predict soil organic carbon. The paper assesses airborne and spaceborne optical sensing.

  2. 2023

    DecodingTrust wins a NeurIPS award

    In 2023, he coauthors DecodingTrust, a trustworthiness evaluation of GPT-4 and GPT-3.5 that examines toxicity, stereotype bias, adversarial robustness, privacy, machine ethics and fairness. The paper wins the NeurIPS Outstanding Paper Award in the Datasets and Benchmarks Track, and its author list marks him and five coauthors as equal contributors.

  3. 2023

    A fellowship for crop sensing

    In 2023, NCSA awards him a second Fiddler Innovation Fellowship for a Students Pushing Innovation project on cross-scale sensing and deep learning for sustainable agriculture. He coauthors a Remote Sensing of Environment paper on field-level crop residue cover that integrates field photos, airborne hyperspectral imagery and satellite data.

  4. 2024

    Testing GPT-4V on Earth imagery

    In 2024, with Sherrie Wang, he coauthors Good at captioning, bad at counting, a CVPR EarthVision workshop benchmark of GPT-4V on Earth-observation data. It tests scene understanding, localization, counting and change detection, and wins the ICLR ML4RS Workshop best-paper award.

  5. March 2025

    MMDT evaluates multimodal trust

    In March 2025, he coauthors MMDT (Multimodal DecodingTrust), a platform that evaluates multimodal foundation models for safety, hallucination, fairness and bias, privacy, adversarial robustness and out-of-distribution generalization.

  6. May 2025

    An SM at MIT IDSS

    In May 2025, he graduates from the MIT Institute for Data, Systems, and Society with an SM in Social and Engineering Systems. Before MIT, he earns a B.S. in computer science at the University of Illinois Urbana-Champaign and works on remote sensing, sustainable agriculture and trustworthy machine learning; Sherrie Wang advises him at MIT, where he works with the Laboratory for Information and Decision Systems and the Earth Intelligence Lab.

  7. July 2025

    AlphaEarth Foundations

    In July 2025, he coauthors AlphaEarth Foundations, an embedding field model for accurate and efficient global mapping from sparse label data. It combines spatial, temporal and measurement contexts to produce maps and monitoring systems from local to global scales. Google DeepMind releases global annual embedding layers covering 2017 through 2024.

  8. As of October 2026

    Research Engineer at Google DeepMind

    As of October 2026, he is a Research Engineer at Google DeepMind, working on AI for Science.

Career

  1. NowMassachusetts Institute of Technology
  2. NowGoogle DeepMindResearch Engineer

Education

  1. University of Illinois Urbana-ChampaignB.S.

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. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Massachusetts Institute of Technology. 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.
  3. 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.
  4. 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.

Sources and credits

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

  1. publications | Chenhui Zhang danielz.ch. Checked 22 October 2026.
  2. [2306.11698] DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models arxiv.org. Checked 22 October 2026.
  3. Chenhui Zhang danielz.ch. Checked 22 October 2026.
  4. NCSA Awards Fiddler Innovation Fellowships to 10 Illinois Students get-savannah.ncsa.illinois.edu. Checked 22 October 2026.
  5. [2401.17600] Good at captioning, bad at counting: Benchmarking GPT-4V on Earth observation data arxiv.org. Checked 22 October 2026.
  6. Publications – earth intelligence lab earthintelligence.mit.edu. Checked 22 October 2026.
  7. [2503.14827] MMDT: Decoding the Trustworthiness and Safety of Multimodal Foundation Models arxiv.org. Checked 22 October 2026.
  8. The 2025 graduates of IDSS idss.mit.edu. Checked 22 October 2026.
  9. [2507.22291] AlphaEarth Foundations: An embedding field model for accurate and efficient global mapping from sparse label data arxiv.org. Checked 22 October 2026.
  10. AlphaEarth Foundations helps map our planet in unprecedented detail — Google DeepMind deepmind.google. Checked 22 October 2026.