← Back to Intelligence Dossier
Oriol Vinyals

Oriol Vinyals

AI Research Leader

Organization
Google DeepMind

Position
VP of Research & Gemini Co-lead, Google DeepMind

πŸ‡ͺπŸ‡ΈSpanish
h-Index100
Citations335,000
Followers15000
Awards1
Publications5
Companies3

Intelligence Briefing

Co-led the development of Gemini at Google DeepMind, including Gemini 2.5 and Gemini 3. Created the AlphaStar agent that defeated professional StarCraft II players. Pioneer of sequence-to-sequence learning β€” his 2014 paper "Sequence to Sequence Learning with Neural Networks" won the NeurIPS Test of Time Award (2024) with over 30,000 citations. Over 335,000 total citations. Joined Google in 2013 and rose to VP of Research by 2023.

Expertise
Natural Language ProcessingSequence-to-Sequence ModelsMultimodal AIGame AINLP
Education

BS, Mathematics and Telecommunication Engineering — Universitat Politècnica de Catalunya

MS, Computer Science β€” University of California, San Diego

PhD, Electrical Engineering and Computer Sciences β€” University of California, Berkeley

Operational History

2026

Gemini 3 Release

Co-led the release of Gemini 3, further advancing AI technology.

research
2025

Gemini 2.5 Release

Co-led the release of Gemini 2.5, enhancing AI capabilities.

research
2024

NeurIPS Test of Time Award

Received the NeurIPS Test of Time Award for his influential paper on sequence-to-sequence learning.

award
2023

Promoted to VP of Research

Promoted to Vice President of Research at Google DeepMind.

career
2014

Published Key Paper

Published 'Sequence to Sequence Learning with Neural Networks', which became foundational in NLP.

research
2013

Joined Google

Began working at Google, contributing to various AI research projects.

career

AGI Position Assessment

Risk Level
LOW
MODERATE
HIGH
CRITICAL
Predicted AGI Timeline

Unknown

Believes in responsible development. Stated there are "no walls in sight" for model capability, emphasizing the need for careful scaling.

Safety Approach

Believes in responsible development. Stated there are "no walls in sight" for model capability, emphasizing the need for careful scaling.

Intercepted Communications

β€œThere are no walls in sight for model capability; we need to be careful about scaling.”

Interview2023-05-15AI Safety

β€œThe future of AI is multimodal; integrating different types of data will be key.”

Conference Talk2023-09-10Multimodal AI

β€œAlphaStar demonstrated that AI can compete at the highest levels in complex games.”

Research Paper2020-01-20Game AI

β€œSequence-to-sequence learning has transformed how we approach NLP tasks.”

Keynote Speech2024-12-01NLP

β€œResponsible AI development is crucial for the future of technology.”

Panel Discussion2025-03-30AI Ethics

Research Output

2020s1
2010s4

Gemini: A New Era in AI

2025

AI Journal

Overview of the Gemini project and its implications.

AlphaStar: Mastering the Real-Time Strategy Game StarCraft II

2019

Nature

Demonstrated AI's capability in complex gaming environments.

w/ Oriol Vinyals, Nicolas Heess, Martin G. Bellemare, Michael Lanctot, Kevin Waugh, Vladimir Kulyukin, Michael Johanson

Pointer Networks

2015

NeurIPS

w/ Oriol Vinyals, Meire Fortunato, Nando de Freitas

Show and Tell: A Neural Image Caption Generator

2015

CVPR

Introduced a novel approach to image captioning.

w/ Oriol Vinyals, Alexander Toshev, Samy Bengio, Google Inc.

Sequence to Sequence Learning with Neural Networks

2014

NeurIPS

Foundational paper in NLP.

30,000 citationsw/ Ilya Sutskever, Quoc V. Le

Field Intelligence

The Future of AI: Multimodal Approaches

●AI Conference 20252025-06-1545 minutes

AI and Ethics: A Responsible Approach

β™ͺTech Podcast2025-02-1030 minutes

Known Associates

Organizational Affiliations

Current

Google DeepMind

Vice President

2013 - Present

Former

Google Brain

Research Scientist

2013 - 2020

Microsoft Research

Research Intern

2012

Commendations

2024

NeurIPS Test of Time Award

NeurIPS

Awarded for the influential paper on sequence-to-sequence learning.

Source Material

Dossier last updated: 2026-03-04