Post-doctoral position in Mathematical Statistics
Topics: generative modeling; high-dimensional robust statistics; optimization
Requirements: a PhD is Statistics, Machine Learning
Duration: 12 months (with possibility to extend)
Gross salary: 3200€ brut mensuel
Work place: L2S, CentraleSupélec, Paris-Saclay University.
Address: 3 rue Joliot Curie, 91190 Gif-sur-Yvette, France.
Contacts: arshak.minasyan@centralesupelec.fr ; avetik.karagulyan@centralesupelec.fr
Project description
The future postdoc, together with the supervisors, will work on a project that lies in the intersection of high-dimensional statistics and optimization. The general aim of the project is to establish optimal (in the minimax sense) learning rates for a given statistical model. The topics include but are not limited to generative modeling, sampling and diffusion models, as well as computational and statistical aspects of robust statistics with possible extension to the online setting.
Upon the scientific interests of the candidate, he/she will work on one of the described topics above.
Desired experience: We seek a motivated researcher with a strong mathematical background and publication record. A PhD in Statistics, Machine Learning or Optimization is required. Coding experience in PyTorch, Tensorflow, Hugging Face or JAX is considered a plus. This position is for young researchers who aim to apply for permanent academic positions next year. In particular, the future postdoc is encouraged to prepare an application for the CNRS concours of Chargé de Recherche and possible future openings at the L2S lab.
The position is available starting September 2025. A National Security clearance is needed, and it can require approximately 2 months.
Requested documents for application: A resume, a motivation letter, and a list of two people who may be contacted for reference letters. A single PDF file containing the requested documents should be sent to arshak.minasyan@centralesupelec.fr and avetik.karagulyan@centralesupelec.fr no later than 31 mars 2025.
