Co-founder of Adaption
A researcher in model efficiency and multilingual AI who led Cohere's research lab and co-founded Adaption; she grew up across Southern and East Africa.
Sara Hooker is a computer scientist known for her work on model efficiency at scale, large language models, and algorithmic bias and fairness. Born in Dublin, Ireland, she moved with her parents to Lesotho at age four and grew up in South Africa, Mozambique, Lesotho, Eswatini and Kenya until she was 19. She earned a PhD in computer science from Mila, the Quebec AI Institute, and in 2014 founded Delta Analytics, which builds technical capacity for non-profits.
She joined Google Brain in 2017, where she worked on model interpretability and pioneered the idea of the Hardware Lottery, and was a founding member of Google's first AI research office in Accra, Ghana. In 2022 she joined Cohere to lead its research arm, Cohere Labs (formerly Cohere For AI), which launched the Aya project, a collaboration of more than 3,000 researchers building a multilingual model and dataset covering 101 languages, and the Aya Expanse models in 2024. In 2025 she co-founded Adaption with Sudip Roy to build adaptable, efficient AI systems. Fortune listed her among AI's top innovators in 2023 and TIME named her to its 2024 list of the most influential people in AI.
As head of Cohere's research lab, led the Aya project, a collaboration of more than 3,000 researchers to create a multilingual model and dataset covering 101 languages, followed by the Aya Expanse models in 2024.
At Google Brain, pioneered the concept of the hardware lottery, which describes how hardware constraints shape the direction of AI research.
Co-founded Adaption in 2025 with Sudip Roy to develop AI systems capable of continuous real-time learning and efficient adaptation.
Founded Delta Analytics in 2014, which develops technical capacity for non-profits.