Oriol Vinyals

Oriol Vinyals is a Catalan-Spanish machine-learning researcher and a senior leader at Google DeepMind, where he is VP of Research and a technical co-lead of the Gemini family of models. Trained in mathematics and telecommunications engineering in Barcelona before earning a PhD at UC Berkeley, he co-invented the influential sequence-to-sequence learning framework with Ilya Sutskever and Quoc Le — a cornerstone of modern translation and language systems.

Vinyals led DeepMind's AlphaStar project, which in 2019 became the first AI to reach Grandmaster level at StarCraft II, and has contributed to landmark systems across language, vision and reinforcement learning; he was named to MIT Technology Review's “35 Innovators Under 35.” His primary focus remains research on Google's flagship models. Across his career he has co-authored some of deep learning's most-cited papers and helped shape several generations of DeepMind's flagship systems.

Vinyals is also a selective angel investor in technically distinctive AI companies. Dealroom tracks his participation in Perplexity and the computer-vision data platform V7. Together with his Gemini leadership and deep-learning pedigree, that activity explains his recurrence on frontier-AI cap tables alongside other researcher-angels. His broader angel activity has also extended to companies such as the AI music-generation startup Udio.