previous
Population genetics and demographic inference
Genomic data often look similar under very different demographic stories, which makes it hard to tell population-size change apart from population structure.
Toulouse, France
Applied Mathematician and AI Researcher
I develop mathematical and machine-learning methods for extracting reliable insight from complex data, with applications spanning population genetics, language, healthcare, and signal processing.

I am an applied mathematician and AI researcher based in Toulouse, France. My early research focused on probabilistic models for reconstructing demographic history from genomic data, with particular attention to population structure, coalescent processes, and the limits of statistical inference. Since then, I have worked across natural-language processing, medical imaging, signal processing, generative models, and AI agents, both in academic collaborations and in industry. I am especially interested in methods that combine mathematical structure, statistical reasoning, and machine learning to produce systems that are scientifically meaningful and practically useful. Alongside my research and engineering work, I teach, mentor student research, and have been president of Toulouse Data Science since 2024.
previous
Genomic data often look similar under very different demographic stories, which makes it hard to tell population-size change apart from population structure.
current
Large text collections are difficult to summarise in a way that remains interpretable and faithful to the underlying corpus.
current
Clinical and scientific images are high-dimensional, sensitive, and often incomplete, which raises questions of representation, uncertainty, and privacy-aware use.
current
Learned components are often most useful when combined with classical models, filters, or tools rather than used in isolation.
Research software and studies
Research on separating population-size variation from population structure using coalescent models and the inverse instantaneous coalescence rate.
Researcher and co-author · INSA Toulouse / Institut de Mathématiques de Toulouse
Coalescent theory · IICR · Identifiability · Simulation
Academic collaborations and mentoring
Ongoing work on comparing transformer-based embeddings and classical topic models for interpretable analysis of large text corpora.
Researcher and mentor
Topic modelling · Transformer embeddings · Coherence measures · Representation learning
Selected applied research
Applied research on decision support, imaging and clinical data workflows, and privacy-aware deployment in healthcare settings.
Applied AI researcher · Torus
Medical imaging · Decision support · Privacy-aware workflows · Generative models
2019
Simona Grusea, , Didier Pinchon, Lounès Chikhi, Simon Boitard, Olivier Mazet
Journal of Mathematical Biology, 78(1), 189-224, 2019
@article{grusea2019coalescence,
title = {Coalescence times for three genes provide sufficient information to distinguish population structure from population size changes},
author = {Grusea, Simona and Rodr{\'i}guez, Willy and Pinchon, Didier and Chikhi, Loun{\`e}s and Boitard, Simon and Mazet, Olivier},
journal = {Journal of Mathematical Biology},
volume = {78},
number = {1},
pages = {189--224},
year = {2019},
doi = {10.1007/s00285-018-1272-4}
}2018
, Olivier Mazet, Simona Grusea, Armando Arredondo, Josué M. Corujo, Simon Boitard, Lounès Chikhi
Heredity, 121(6), 663-678, 2018
@article{rodriguez2018nssc,
title = {The IICR and the non-stationary structured coalescent: towards demographic inference with arbitrary changes in population structure},
author = {Rodr{\'i}guez, Willy and Mazet, Olivier and Grusea, Simona and Arredondo, Armando and Corujo, Josu{\'e} M. and Boitard, Simon and Chikhi, Loun{\`e}s},
journal = {Heredity},
volume = {121},
number = {6},
pages = {663--678},
year = {2018},
doi = {10.1038/s41437-018-0148-0}
}2018
Lounès Chikhi, , Simona Grusea, Patrícia Santos, Simon Boitard, Olivier Mazet
Heredity, 120(1), 13-24, 2018
@article{chikhi2018iicr,
title = {The IICR (inverse instantaneous coalescence rate) as a summary of genomic diversity: insights into demographic inference and model choice},
author = {Chikhi, Loun{\`e}s and Rodr{\'i}guez, Willy and Grusea, Simona and Santos, Patr{\'i}cia and Boitard, Simon and Mazet, Olivier},
journal = {Heredity},
volume = {120},
number = {1},
pages = {13--24},
year = {2018},
doi = {10.1038/s41437-017-0005-6}
}2016
Simon Boitard, , Flora Jay, Stefano Mona, Frédéric Austerlitz
PLOS Genetics, 12(3), article e1005877, 2016
DOI ·
@article{boitard2016popsize,
title = {Inferring Population Size History from Large Samples of Genome-Wide Molecular Data - An Approximate Bayesian Computation Approach},
author = {Boitard, Simon and Rodr{\'i}guez, Willy and Jay, Flora and Mona, Stefano and Austerlitz, Fr{\'e}d{\'e}ric},
journal = {PLOS Genetics},
volume = {12},
number = {3},
pages = {e1005877},
year = {2016},
doi = {10.1371/journal.pgen.1005877}
}2015
Olivier Mazet, , Lounès Chikhi
Theoretical Population Biology, 104, 46-58, 2015
DOI ·
@article{mazet2015demographic,
title = {Demographic inference using genetic data from a single individual: Separating population size variation from population structure},
author = {Mazet, Olivier and Rodr{\'i}guez, Willy and Chikhi, Loun{\`e}s},
journal = {Theoretical Population Biology},
volume = {104},
pages = {46--58},
year = {2015},
doi = {10.1016/j.tpb.2015.06.003}
}I teach and mentor in applied mathematics, probability, machine learning, and NLP. Public course details are listed only when titles, dates, and roles have been confirmed.
MIT PRIMES · mentor · 2025–2026
2025-11-27
Toulouse Data Science
2016-06-20
PhD defence
I welcome conversations about research collaboration, academic teaching, mentoring, and scientific talks.