Noam Shazeer is an American computer scientist known as one of the eight co-authors of the 2017 paper that introduced the Transformer (architecture) architecture, and as the co-founder of the chatbot company Character.AI.
Career at Google
Shazeer worked at Google for many years before the transformer era, contributing to large-scale systems including parts of Google's spelling correction and early Mixture of experts research, an architecture he helped popularize as a way to scale model capacity without proportionally scaling compute cost. In 2017, he was a co-author of "Attention Is All You Need," the paper that introduced the transformer, which replaced recurrence with an self-attention mechanism and became the architectural basis for nearly all major large language models developed since, including BERT and GPT-3. All eight authors of the paper eventually left Google.
Character.AI
In 2021, Shazeer left Google to co-found Character.AI, a company that let users create and converse with customizable AI personas, built on large language models the company trained itself. Character.AI grew rapidly, reaching tens of millions of users, and became one of the most prominent examples of AI companion products, but it also drew scrutiny and lawsuits over the safety of its interactions with younger users and a widely reported case involving user harm.
Return to Google
In August 2024, Google announced a deal in which it paid to license Character.AI's technology and hired much of its leadership, including Shazeer, who returned to Google as a senior executive within the Google DeepMind organization, contributing to the development of the Gemini model family. The arrangement, structured as a licensing deal plus talent hire rather than a formal acquisition, was widely seen as part of a pattern of large AI labs re-absorbing founders of smaller AI companies, a structure also used in deals involving Inflection AI and Microsoft.
Influence
As a co-author of the transformer paper, Shazeer is credited as one of the architects of the technical foundation underlying the entire generative AI industry that emerged after 2022, and his career arc from Google to a consumer AI startup and back is frequently cited as illustrative of how concentrated the frontier AI talent market became during the generative AI boom.