Ahmet Üstün

Researcher, Cohere For AI

Joined

January 2022

Names

Ahmet Üstün (Preferred)

Emails

****@rug.nl (Confirmed)

****@gmail.com (Confirmed)

****@cohere.com (Confirmed)

Personal Links

Career & Education History

Researcher

Cohere For AI (cohere.com)

2023 – Present

PhD student

University of Groningen (rug.nl)

2018 – 2023

MS student

Middle East Technical University (metu.edu.tr)

2015 – 2017

Undergrad student

Middle East Technical University (metu.edu.tr)

2007 – 2014

Advisors, Relations & Conflicts

Coworker

Sara Hooker

2022 – Present

PhD Advisor

Gertjan van Noord

2018 – Present

PhD Advisor

Gosse Bouma

2018 – Present

PhD Advisor

Arianna Bisazza

2018 – Present

Expertise

Efficient Deep Learning

Present

Mixture of Experts

Present

Multilingual LLMs

Present

Multimodal LLMs

Present

Publications

Turning Speech Language Models into Multilingual Listeners
Tolúlopé Ògúnrèmí, Dan Jurafsky, Christopher D Manning, Ahmet Üstün, Martijn Bartelds
- Submitted to ICLR 2026
- Readers: Everyone
Aya Vision: Advancing the Frontier of Multilingual Multimodality
Saurabh Dash, Yiyang Nan, John Dang, Arash Ahmadian, Shivalika Singh, Madeline Smith, Bharat Venkitesh, Vladyslav Shmyhlo, Viraat Aryabumi, Walter Beller-Morales, Jeremy Pekmez, Jason Ozuzu, Pierre Harvey Richemond, Acyr Locatelli, Nicholas Frosst, Phil Blunsom, Aidan Gomez, Ivan Zhang, Marzieh Fadaee, Manoj Govindassamy et al. (5 additional authors not shown)
- Submitted to ICLR 2026
- Readers: Everyone
EAGER: Entropy-Aware GEneRation for Adaptive Inference-Time Scaling
Daniel Scalena, Leonidas Zotos, Elisabetta Fersini, Malvina Nissim, Ahmet Üstün
- Submitted to ICLR 2026
- Readers: Everyone
The Disparate Impacts of Speculative Decoding
Jameson Sandler, Ahmet Üstün, Marco Romanelli, Sara Hooker, Ferdinando Fioretto
- Submitted to ICLR 2026
- Readers: Everyone
One Tokenizer To Rule Them All: Emergent Language Plasticity via Multilingual Tokenizers
Diana Abagyan, Alejandro R. Salamanca, Andres Felipe Cruz-Salinas, Kris Cao, Hangyu Lin, Acyr Locatelli, Marzieh Fadaee, Ahmet Üstün, Sara Hooker
- Submitted to ICLR 2026
- Readers: Everyone
Treasure Hunt: Real-time Targeting of the Long Tail using Training-Time Markers
Daniel D'souza, Julia Kreutzer, Adrien Morisot, Ahmet Üstün, Sara Hooker
- NeurIPS 2025 poster
- Readers: Everyone
The Leaderboard Illusion
Shivalika Singh, Yiyang Nan, Alex Wang, Daniel D'souza, Sayash Kapoor, Ahmet Üstün, Sanmi Koyejo, Yuntian Deng, Shayne Longpre, Noah A. Smith, Beyza Ermis, Marzieh Fadaee, Sara Hooker
- NeurIPS 2025 Datasets and Benchmarks Track poster
- Readers: Everyone
Aya Vision: Advancing the Frontier of Multilingual Multimodality
Saurabh Dash, Yiyang Nan, John Dang, Arash Ahmadian, Shivalika Singh, Madeline Smith, Bharat Venkitesh, Vladyslav Shmyhlo, Viraat Aryabumi, Walter Beller-Morales, Jeremy Pekmez, Jason Ozuzu, Pierre Harvey Richemond, Acyr Locatelli, Nicholas Frosst, Phil Blunsom, Aidan Gomez, Ivan Zhang, Marzieh Fadaee, Manoj Govindassamy et al. (5 additional authors not shown)
- Submitted to NeurIPS 2025
- Readers: Everyone
Nexus: Specialization meets Adaptability for Efficiently Training Mixture of Experts
Nikolas Gritsch, Qizhen Zhang, Acyr Locatelli, Sara Hooker, Ahmet Üstün
- AFM 2024 Poster
- Readers: Everyone
Seeded LoRA: Collaborative Fine-Tuning Through Seed Initialization of Adapters
Alejandro R. Salamanca, Ahmet Üstün, Nicki Skafte Detlefsen, Tim Dettmers
- Submitted to ICLR 2025
- Readers: Everyone

View all 65 publications

Co-Authors

View all 139 co-authors