# Oriol Vinyals

Oriol Vinyals (born 1983) is a Spanish machine learning researcher known for co-inventing the seq2seq model, leading AlphaStar at DeepMind, and serving as technical lead on Gemini. He co-founded Discovery Loop in 2026.

Oriol Vinyals (born 1983) is a Spanish machine learning researcher who has spent most of his career at [DeepMind](https://www.wikiprompt.org/wiki/google-deepmind). He is known for co-inventing the sequence-to-sequence (seq2seq) model for machine translation, leading the AlphaStar research group, and serving as technical lead on the Gemini project. In 2026, he co-founded the startup Discovery Loop with several colleagues.

## Education and early career

Vinyals was born in Sabadell, Catalonia, Spain. He studied mathematics and telecommunication engineering at the Universitat Politècnica de Catalunya (UPC). He then moved to the United States, earning a Master's degree in computer science from the University of California, San Diego and later from the University of California, Berkeley. He received his PhD in 2013 from Berkeley's Department of Electrical Engineering and Computer Sciences under the supervision of Nelson Morgan.

## Seq2seq and early contributions

While working at [DeepMind](https://www.wikiprompt.org/wiki/google-deepmind) and collaborating with researchers at [OpenAI](https://www.wikiprompt.org/wiki/openai), Vinyals co-invented the seq2seq model for machine translation along with Ilya Sutskever and Quoc Viet Le. This model, which uses a [neural network](https://www.wikiprompt.org/wiki/neural-network) to map sequences to sequences, became a foundational technique in [machine learning](https://www.wikiprompt.org/wiki/machine-learning) and underpinned many later advances in [large language models](https://www.wikiprompt.org/wiki/large-language-model). The approach allowed for more flexible and accurate translation and text generation compared to earlier methods.

## AlphaStar and research leadership

Vinyals led the AlphaStar research group at [DeepMind](https://www.wikiprompt.org/wiki/google-deepmind), which applied [artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence) to computer games such as StarCraft II. The project demonstrated significant progress in [reinforcement learning](https://www.wikiprompt.org/wiki/reinforcement-learning) and real-time strategy gameplay, achieving grandmaster level performance in the game. His research at DeepMind has been regularly featured in mainstream media, highlighting breakthroughs in [deep learning](https://www.wikiprompt.org/wiki/deep-learning) and [generative AI](https://www.wikiprompt.org/wiki/generative-ai).

In 2016, MIT Technology Review named Vinyals one of the 35 most innovative young people under 35. By 2022, he was a principal research scientist at DeepMind, continuing to contribute to the field of [artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence).

## Gemini and Discovery Loop

As of August 2024, Vinyals was technical lead on Gemini, a major [large language model](https://www.wikiprompt.org/wiki/large-language-model) project at [DeepMind](https://www.wikiprompt.org/wiki/google-deepmind), working alongside Noam Shazeer and Jeff Dean. In 2025, he received an honorary doctoral degree from his home university, UPC.

On August 5, 2026, Vinyals announced he was leaving Gemini to start Discovery Loop, a new venture, together with Jeff Dean, Quoc Le, and Sanjay Ghemawat. The company focuses on advancing research in [artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence) and [machine learning](https://www.wikiprompt.org/wiki/machine-learning).

## See also

- [Ilya Sutskever](https://www.wikiprompt.org/wiki/ilya-sutskever)

---
Source: https://www.wikiprompt.org/wiki/oriol-vinyals
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-09-05T13:26:41.974745+00:00