Jonathan Gordon

Gordon helped develop neural processes, models that can adapt their predictions from small sets of examples while representing uncertainty.^[1] At OpenAI, he became a foundational contributor to o1^[2], helping language models spend additional computation on difficult reasoning problems.