The EPFL AI Center brings together outstanding early-career researchers who collaborate with EPFL faculty across diverse scientific domains to advance the capabilities and applications of AI.
With a focus on excellence and collaborative research, our fellows work on forefront AI projects while contributing to the Center’s community and strategic initiatives.
Below, meet the fellows and researchers currently shaping the future of AI at the EPFL AI Center, including researchers supported through the AI Fellowship Programme 
.
Multimodality, 3D & Spatial Intelligence
Medical AI, multimodal & active learning
Linus develops efficient and trustworthy algorithms for multimodal healthcare data. His research focuses on transfer learning and active learning: reusing knowledge from related tasks, selectively acquiring new informative data, and applying these methods to novel precision healthcare problems.
EPFL Hosts: Charlotte Bunne & Bart Deplancke
Medical AI, evaluation & global health
AI, optimization & energy efficiency
Arto develops optimization methods to make the training of foundation models faster and more resource-efficient. His research addresses major bottlenecks in distributed training, developing methods that reduce communication costs and allow faster devices to continue working without waiting for slower ones.
EPFL Hosts: Martin Jaggi & Volkan Cevher
Safe and Robust AI
Andi studies how groups of large language models can be organised to solve complex tasks more effectively. His research focuses on multi-agent orchestration: designing systems in which specialized agents should interact, collaborate and take actions to achieve the best possible outcome.
EPFL Hosts: Maryam Kamgarpour & Volkan Cevher
GenAI, verification & interactive proofs
Orr develops “Self-Proving Models”: generative models that can formally prove their answers are correct. His research builds on interactive proof systems, where an AI generates an answer and interacts with a verification algorithm until the verifier is mathematically certain the answer is correct.
EPFL Hosts: Nicolas Flammarion, Thomas Bourgeat, Lenaïc Chizat, Viktor Kuncak.
AI for Health & Humanitarian response
NLP, ML and cognitive neuroscience
Badr works at the intersection of natural language processing, machine learning, and cognitive neuroscience. His research explores how large language models can be adapted to better capture the neural and behavioral responses of unseen individuals or populations, with a particular focus on cognitive disorders.
EPFL Hosts: Antoine Bosselut & Martin Schrimpf
AI architecture and efficiency
Generative models & diffusion methods
DL & statistical physics of computation
Generative models & molecule design
LLMs, AI for Education & Interpretability
ML, CV & multimodal models
DL models & model of perception
Scaling methods & LLMs optimization
ML & Computational Neuroscience
NLP, ML and AI for scientific discovery
Mete studies the limitations of large language models as tools for scientific discovery. His research focuses particularly on scientific creativity, exploring how LLMs generate promising hypotheses, develop new ideas and make connections beyond reproducing knowledge, and how their capacity for scientific reasoning and ideation can be improved.
EPFL Host: Antoine Bosselut
Representation learning & drug discovery
Multi-agent learning & game theory
LLMs Post-training & Medical AI
Models pretraining & interpretability
Computational chemistry & biology
Agentic AI & AI for scientific discovery
Siba develops AI agents to support scientific discovery from single-cell biological data. His research focuses on training and evaluating LLM-based agents that can explore complex datasets, propose hypotheses and run computational analyses, with particular attention to reliability and reproducibility.
EPFL Host: Maria Brbic
Interpretability & training dynamics
Pretraining, AI safety & alignment
Computational chemistry & biology
Foundation model & spatial proteomics
Computational biology & protein design
Optimization & Data efficiency
ML & high dimensional statistics
Model scaling & multimodality
Generative & geometric deep learning
Theory of Machine Learning
Pretraining, learning & generalization
Generative models and AI for drugs
Judge models & post-training
AI interpretability
Fundamental AI methods & behaviour
ML, biology & immune cell variation
LLM training & model development
GenAI for drug discovery
Natural Language Processing
Reasoning & language modelling
Multilinguality & modularity in models
