PeopleofSuperintelligence

Lab leader

Michael Isard

Also known as Michael Acheson Isard

Michael Isard is a computer vision and systems researcher whose work spans visual tracking, distributed systems, and machine learning infrastructure. He earned a D.Phil. in computer vision from Oxford in 1998 and helped introduce particle filters through Condensation. At Microsoft Research, he worked on distributed systems such as Dryad. He co-authored TensorFlow, received the 2017 Longuet-Higgins Prize for an object retrieval paper, and co-led Gemini 1.0 infrastructure; the Royal Society elected him a Fellow in 2023. As of October 2026, he is a Google DeepMind research scientist who co-leads a team building infrastructure used to train Google's largest machine learning models.

Works
135
Citations
49,244
h-index
54

Publication numbers from OpenAlex, as of 14 October 2026.

Photo of Michael Isard

Timeline

  1. 1998

    Oxford and Condensation

    In 1998, Michael Acheson Isard earns a D.Phil. in computer vision from Oxford and develops the Condensation algorithm for visual tracking. His work helps introduce particle filters to computer vision.

  2. 2002–2007

    Microsoft Research and Dryad

    In 2002, he works at Microsoft Research Silicon Valley on visual tracking and sequential filtering. In 2007, he co-authors Dryad, a distributed execution engine for data-processing programs that scales to data centers with thousands of computers.

  3. 2015

    TensorFlow goes open source

    In November 2015, Google releases the TensorFlow API and reference implementation as open source; Isard co-authors the report describing its large-scale machine learning system.

  4. 2017

    A Longuet-Higgins Prize

    In 2017, the Longuet-Higgins Prize goes to the 2007 CVPR paper he co-authors with James Philbin, Ondrej Chum, Josef Sivic, and Andrew Zisserman, “Object Retrieval with Large Vocabularies and Fast Spatial Matching.” The system returns ranked images that contain a query object from a large corpus.

  5. 2023

    Royal Society Fellowship

    The Royal Society elects him a Fellow in 2023.

  6. 2025

    Gemini infrastructure co-lead

    In 2025, he co-leads infrastructure for Gemini 1.0, Google's multimodal model family.

  7. April 2026

    Advising on Decoupled DiLoCo

    In April 2026, he provides additional support and advice to Google DeepMind's Decoupled DiLoCo work on distributed AI training.

  8. As of October 2026

    Google DeepMind research scientist

    As of October 2026, he is a Research Scientist at Google DeepMind and co-leads a team building infrastructure used to train Google's largest and most capable machine learning models.

Sources and credits

Photo: research.google; Michael Isard; creator not stated. License: Not stated by source. Source: research.google. Cropped to a square and resized. Ask us to remove or replace this photo.

  1. Michael Acheson Isard | Innovators Under 35 innovatorsunder35.com. Checked 14 October 2026.
  2. Faculty Summit 2006 - Microsoft Research microsoft.com. Checked 14 October 2026.
  3. Dryad: Distributed Data-parallel Programs from Sequential Building Blocks microsoft.com. Checked 14 October 2026.
  4. TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems arxiv.org. Checked 14 October 2026.
  5. Two prizes awarded to Josef Šivic in 2017 ciirc.cvut.cz. Checked 14 October 2026.
  6. Object Retrieval with Large Vocabularies and Fast Spatial Matching microsoft.com. Checked 14 October 2026.
  7. Exceptional scientists elected as Fellows of the Royal Society royalsociety.org. Checked 14 October 2026.
  8. Gemini: A Family of Highly Capable Multimodal Models arxiv.org. Checked 14 October 2026.
  9. Decoupled DiLoCo: Resilient, Distributed AI Training at Scale deepmind.google. Checked 14 October 2026.
  10. Dr Michael Isard FRS | Royal Society Fellow royalsociety.org. Checked 14 October 2026.