Researcher

Kory Wallace Mathewson

Kory Wallace Mathewson is a researcher whose work spans reinforcement learning, robotics, language models and creative tools. He co-authored research on human feedback for robotic learning and built an AI improviser for live shows with human actors. His language-model projects include Story Centaur for writers and Dramatron for scripts. He is a Staff Research Scientist at Google DeepMind, studies how humans and machines create together, and launched the Veo video models and Flow filmmaking tool.

Timeline

  1. 2016–2017

    Robot learning and AI improv

    In 2016, he coauthors a study on training a robotic agent with actor-critic reinforcement learning and simultaneous human control and feedback. In 2017, he and Piotr Mirowski build an AI actor for live improvised theatre; their paper describes 30 shows and a Turing test for the audience.

  2. By 2019

    Engineering and a PhD

    Before his PhD, he earns a bachelor's degree in electrical engineering and a master's degree in biomedical engineering. In November 2019, he completes a PhD in Computing Science at the University of Alberta.

  3. By 2019

    Research groups at UAlberta

    By 2019, he works with the Alberta Machine Intelligence Institute (Amii), the Reinforcement Learning and Artificial Intelligence (RLAI) group, and the Bionic Limbs for Improved Natural Control (BLINC) lab.

  4. 2019

    DeepMind in Montreal

    In November 2019, he moves to Montreal to work as a research scientist with DeepMind.

  5. 2021

    Story Centaur for writers

    In 2021, he coauthors Story Centaur, an interface that helps creative writers prototype few-shot models and compose their own co-creation tools.

  6. 2023

    Dramatron co-writing scripts

    In 2023, he coauthors Dramatron, which applies language models hierarchically and uses prompt chaining to generate coherent scripts and screenplays.

  7. 2025

    Technical direction for ANCESTRA

    In June 2025, he serves as Technical Director for ANCESTRA, which premieres at the Tribeca Film Festival. The short pairs live-action scenes with video generated by Veo.

  8. As of October 2026

    Google DeepMind staff scientist

    As of October 2026, he is a Staff Research Scientist at Google DeepMind and studies how humans and machines create together. He launches the Veo video models and the Flow filmmaking tool, and contributes to Gemini Omni.

Career

  1. Since 2019Google DeepMindResearch Scientist

Education

  1. 2019University of AlbertaPhD in Computing Science

Sources and credits

Photo: We have no published photo of Kory Wallace Mathewson yet. Send us a free-licensed photo.

  1. Simultaneous Control and Human Feedback in the Training of a Robotic Agent with Actor-Critic Reinforcement Learning arxiv.org. Checked 21 October 2026.
  2. Improvised Comedy as a Turing Test arxiv.org. Checked 21 October 2026.
  3. Improv and AI: Science grad connects comedy and science | Faculty of Science ualberta.ca. Checked 21 October 2026.
  4. Kory Mathewson | Google Research research.google. Checked 21 October 2026.
  5. Co-Writing Screenplays and Theatre Scripts with Language Models: An Evaluation by Industry Professionals deepmind.google. Checked 21 October 2026.
  6. Behind “ANCESTRA”: combining Veo with live-action filmmaking blog.google. Checked 21 October 2026.
  7. Kory Mathewson | Research Scientist korymathewson.com. Checked 21 October 2026.