PeopleofSuperintelligence

Researcher

Noam Shazeer

Noam Shazeer is a researcher and engineer whose work connects language-model design with the systems needed to train large models. He coauthored the 2017 sparse mixture-of-experts paper and the Transformer paper with their research teams. Later papers included Mesh-TensorFlow, T5, question answering from model parameters, and LaMDA. These projects address different parts of the same problem: how to build models that process language at scale. In 2022, he and Daniel de Freitas introduced Character.AI, with a founding team developing conversational characters. His early work was at Google.

Now at
Character.AI
Works
58

Publication numbers from OpenAlex, as of 14 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Photo of Noam Shazeer

Timeline

  1. Early Google work

    Language in useful software

    Shazeer recalls work on spelling suggestions and ad targeting.

  2. 2017

    Experts and attention

    He coauthors sparse mixture-of-experts research and the Transformer paper with separate research teams.

  3. 2018

    Distributed model training

    Mesh-TensorFlow distributes calculations across a mesh of processors.

  4. 2019–2020

    Shared language tasks

    The T5 team uses text-in, text-out tasks; Roberts, Raffel, and Shazeer also study closed-book question answering.

  5. 2022

    Dialogue models

    Shazeer is among LaMDA’s authors, studying dialogue quality, safety, and grounding.

  6. 2022

    Character.AI

    He and Daniel de Freitas introduce Character.AI after its September public beta.

Sources and credits

Photo: Axios HQ event speaker profile. License: Not stated by source. Source: axioshq.com. Cropped to a square and resized. Ask us to remove or replace this photo.

  1. Shazeer, de Freitas and team: Introducing Character, 2022 blog.character.ai. Checked 10 October 2026.
  2. Shazeer and colleagues: sparse mixture of experts, 2017 arxiv.org. Checked 10 October 2026.
  3. Vaswani and colleagues: Attention Is All You Need, 2017 arxiv.org. Checked 10 October 2026.
  4. Shazeer and colleagues: Mesh-TensorFlow, 2018 arxiv.org. Checked 10 October 2026.
  5. Raffel and colleagues: T5, 2019 arxiv.org. Checked 10 October 2026.
  6. Roberts, Raffel and Shazeer: model knowledge, 2020 arxiv.org. Checked 10 October 2026.
  7. Thoppilan and colleagues: LaMDA, 2022 arxiv.org. Checked 10 October 2026.