Eric Moulines is a French mathematician and computer scientist. He is Dean of the Computing and Mathematical Sciences Division and Professor of Machine Learning at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), and he is also Full Professor with EPITA. Across a career spanning more than three decades, he has built scientific profile at the intersection of probability, statistics, optimization, and artificial intelligence, with a particular emphasis on algorithms that are both theoretically grounded and practically useful for complex data.
Eric Moulines earned an engineering degree from École Polytechnique and a Ph.D. in electrical engineering from the École Nationale Supérieure des Télécommunications. He began his research career in speech processing, working on text-to-speech synthesis and waveform processing methods. In 1990, he joined the Signal and Image Processing Department at Télécom ParisTech, where he was appointed full professor in 1996. His early contributions to speech and signal processing were followed by influential work on blind source separation, subspace methods for system identification, adaptive estimation, and time series modeling.
As his research evolved, Eric Moulines became one of the leading figures in computational statistics and Bayesian inference. He made important contributions to hidden Markov models, nonlinear filtering, sequential Monte Carlo methods, Markov chain Monte Carlo, and stochastic approximation. A recurring theme in his work is the rigorous analysis of algorithms used to perform inference in models that are analytically intractable but central to applications in signal processing, machine learning, and data science. His research combines mathematical depth with a strong awareness of computational constraints, especially in settings involving latent variables, partial observations, and high-dimensional data.
In 2015, Moulines moved to the Centre de Mathématiques Appliquées at École Polytechnique as Professor of Statistics. His later work has focused increasingly on high-dimensional statistical inference, Monte Carlo sampling, Bayesian inverse problems, stochastic optimization, federated and distributed statistics, and the theory of algorithms for machine learning. He has also contributed to the analysis of Langevin algorithms, Hamiltonian Monte Carlo, and non-asymptotic convergence bounds, helping to clarify the behavior of sampling and optimization methods that are now widely used in modern artificial intelligence.
At EPITA his research interests include high-dimensional Monte Carlo methods, stochastic optimization, generative models, uncertainty quantification, Bayesian inverse problems, and the control of complex systems. He contributes to the development of a research and education environment that connects mathematical sciences with artificial intelligence, emphasizing the foundations needed to build reliable, scalable, and interpretable AI systems.
Eric Moulines has published extensively, with more than 120 journal articles and more than 300 conference papers in signal processing, computational statistics, applied probability, and machine learning. His scientific influence is reflected not only in his publications, but also in his mentorship, editorial service, and participation in the international statistics and signal processing communities. He has served on the editorial boards of major journals and was Editor-in-Chief of Bernoulli from 2013 to 2016.
His achievements have been recognized by several major distinctions. He received the CNRS Silver Medal in 2010 and the Orange Prize of the French Academy of Sciences in 2011. He is a Fellow of EURASIP and of the Institute of Mathematical Statistics, received the EURASIP Technical Achievement Award in 2020, and was elected to the French Academy of Sciences in 2017.