PeopleofSuperintelligence

Researcher

Aurelia Guy

Aurelia Guy is an AI researcher whose work spans reinforcement learning, interactive agents, and language models. In 2018, she co-authored a random-search method for continuous control, and in 2020 she co-authored Google DeepMind's Imitating Interactive Intelligence on a simulated Playroom for interactive agents. Her 2021 and 2022 papers studied Gopher, retrieval-enhanced models, routed scaling laws, and speech-based task learning. The compute-optimal training paper she co-authored earned a NeurIPS 2022 Outstanding Paper award. By 2024, she contributed to OpenAI's ChatGPT and Whisper API release.

Timeline

  1. 2018

    Random search for reinforcement learning

    In 2018, she co-authors a random-search method for training static linear policies for continuous control, matching state-of-the-art sample efficiency on MuJoCo locomotion tasks.

  2. 2020

    Interactive agents in the Playroom

    In 2020, she co-authors Google DeepMind's Imitating Interactive Intelligence, which introduces a simulated Playroom where agents move objects, speak, follow instructions, and answer questions.

  3. 2021

    Training Gopher at scale

    In 2021, she co-authors Scaling Language Models: Methods, Analysis & Insights from Training Gopher, which evaluates Transformer models up to 280 billion parameters on 152 tasks.

  4. 2022

    Retrieval-Enhanced Transformer

    In 2022, she co-authors Improving Language Models by Retrieving from Trillions of Tokens, which introduces RETRO and conditions language models on retrieved text from a large corpus.

  5. 2022

    Scaling laws for routed models

    In 2022, she co-authors Unified Scaling Laws for Routed Language Models, which derives scaling laws for routing networks that conditionally use a subset of their parameters to process an input.

  6. 2022

    Speech for zero-shot task learning

    In 2022, she co-authors Intra-agent speech permits zero-shot task acquisition, a study that teaches embodied agents to caption images and manipulate new objects with little labeled language data.

  7. 2022

    A NeurIPS Outstanding Paper

    In 2022, she co-authors An empirical analysis of compute-optimal large language model training, which earns a NeurIPS Outstanding Paper award. Its Chinchilla model uses the same compute budget as Gopher with 70 billion parameters.

  8. By 2024

    OpenAI API contributions

    By April 2024, she contributes to OpenAI's launch of the ChatGPT and Whisper APIs.

  1. Google had the most people shaping AI every year from 2014 to 2025Our 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. The United States employs an estimated 48% of people shaping AIWe estimate where the people shaping AI work today, how universities, companies and research institutes split across countries, and how country shares have changed since the 1960s.
  3. 43% of living people shaping AI work outside their country of birthWe estimate that 43% of living people shaping AI work outside their country of birth (520 of 1,198 people with known birth and employment countries).
  4. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionWe 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.

Sources and credits

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

  1. Simple random search provides a competitive approach to reinforcement learning arxiv.org. Checked 20 October 2026.
  2. Imitating Interactive Intelligence deepmind.google. Checked 20 October 2026.
  3. Scaling Language Models: Methods, Analysis & Insights from Training Gopher arxiv.org. Checked 20 October 2026.
  4. Improving Language Models by Retrieving from Trillions of Tokens proceedings.mlr.press. Checked 20 October 2026.
  5. Unified Scaling Laws for Routed Language Models arxiv.org. Checked 20 October 2026.
  6. Intra-agent speech permits zero-shot task acquisition papers.neurips.cc. Checked 20 October 2026.
  7. Announcing the NeurIPS 2022 Awards blog.neurips.cc. Checked 20 October 2026.
  8. An empirical analysis of compute-optimal large language model training papers.nips.cc. Checked 20 October 2026.
  9. Introducing ChatGPT and Whisper APIs web.archive.org. Checked 20 October 2026.