Vincent Cohen-Addad
Vincent Cohen-Addad is a research scientist at Google Research in New York who designs clustering and optimization algorithms and works on reasoning in large AI models. After graduating from École Normale Supérieure in Paris, he won the EATCS Distinguished Dissertation Award 2016 for his Ph.D. thesis, joined CNRS in 2017, and joined Google Research in 2020. In March 2025, he and four coauthors submitted a (2+ε)-approximation algorithm for metric k-median, and in July 2025 he was among the leads and key contributors of the team whose Gemini Deep Think model reached gold-medal level at the International Mathematical Olympiad. He was a joint recipient of the 2026 EATCS Presburger Award. His current work covers inference methods and reinforcement learning algorithms that train foundation models to reason better, and retrieval-augmented generation for products such as AI Overview and AI Mode.
- Now at
- Google Research
- Badges
Timeline
-
2016
A thesis prize from EATCS
He graduates from École Normale Supérieure in Paris and receives the EATCS Distinguished Dissertation Award 2016 for his Ph.D. thesis. He also serves as a Marie Skłodowska-Curie Fellow at BARC, the Basic Algorithms Research Copenhagen group at the University of Copenhagen.
-
2017–2020
CNRS researcher at LIP6
He joins CNRS in France in 2017 as a chargé de recherche based at Sorbonne Université within the LIP6 laboratory, where he stays until 2020. In 2019, he wins the Best Paper Award at SoCG.
-
2020–2022
Joining Google Research
He joins Google Research in 2020. In 2022, he is an invited plenary speaker at APPROX.
-
2024
An ESA invited talk
In 2024, he gives an invited talk at the 32nd European Symposium on Algorithms on correlation clustering, a model used in machine learning and data mining. He reviews how a simple, practical algorithm from 2021 achieves an O(1)-approximation in massively parallel, sublinear, streaming and differentially private models.
-
March 2025
A (2+ε)-approximation for k-median
In March 2025, he and Fabrizio Grandoni, Euiwoong Lee, Chris Schwiegelshohn and Ola Svensson submit a (2+ε)-approximation algorithm for metric k-median, which improves the previous best factor of 2.613.
-
July 2025
Gemini Deep Think at the IMO
In July 2025, an advanced version of Gemini Deep Think solves five of the six International Mathematical Olympiad problems perfectly and earns 35 points, a gold-medal level result. He is among the leads and key contributors of the IMO 2025 team at Google DeepMind.
-
2026
A joint Presburger Award
He becomes a joint recipient of the 2026 EATCS Presburger Award.
-
As of October 2026
Google Research in New York
He is a Research Scientist at Google Research in New York. His work covers inference methods and reinforcement learning algorithms for training foundation models to reason better, optimization algorithms for data analysis and machine learning, and retrieval-augmented generation for products such as AI Overview and AI Mode.
Career
- Since 2020Google ResearchResearch Scientist
- 2017–2020Sorbonne UniversityChargé de recherche
- 2016University of CopenhagenMarie Skłodowska-Curie Fellow
Education
Sources and credits
Photo: We have no published photo of Vincent Cohen-Addad yet. Send us a free-licensed photo.
- BARC talk by Vincent Cohen-Addad – University of Copenhagen
- Cohen-Addad, Vincent (19..-....) – IdRef
- Recent Progress on Correlation Clustering: From Local Algorithms to Better Approximation Algorithms and Back (Invited Talk)
- A (2+ε)-Approximation Algorithm for Metric k-Median
- Advanced version of Gemini with Deep Think officially achieves gold-medal standard at the International Mathematical Olympiad
- Vincent Cohen-Addad - Research Scientist




