Rémy Hosseinkhan-Boucher

Rémy Hosseinkhan-Boucher

Postdoc @ CMAP École Polytechnique
PhD in Reinforcement Learning @ Paris-Saclay University

Hello! I am a Post-doctoral researcher at Centre de Mathématiques Appliquées de l’École Polytechnique (CMAP) in the HPC@Maths team.

I recently completed my PhD in Reinforcement Learning at Université Paris-Saclay, where I worked on Learning-based Control of Dynamical Systems with applications in Computational Fluid Dynamics.
My research was supervised by A. Vilnat, O. Semeraro, and L. Mathelin and my PhD thesis is available here.

Before, I was a research intern at Inria TAU, where I worked on Learning-based methods for Stiff Differential Equations: Koopman operator theory and Physics-Informed Neural Networks under the supervision of M-A. Bucci, T. Faney, C. Mehl and M. Schoenauer.
I have also been a research intern in several R&D departments such as Capital Fund Management (CFM) (Anomaly Detection in Time Series), BNP Paribas Real Estate (Machine Learning for Real Estate Market Analysis), and Luxurynsight (Deep Learning for Natural Language Processing).

I hold an MSc in Artificial Intelligence, Systems, and Data, an MSc in Statistics and Financial Mathematics and a BSc in Applied Mathematics (probability and statistics) from Université Paris Dauphine - PSL.

Research Interests

  • Reinforcement Learning
  • Learning-based Control
  • Information Theory
  • Physics-informed Machine Learning

Publications

  • Learning non-Markovian Dynamical Systems with Signature-based Encoders
    E. Pradeleix, R. Hosseinkhan-Boucher, A. Shilova, O. Semeraro, L. Mathelin
    Proceedings of the ECAI Workshop on “Machine Learning Meets Differential Equations: From Theory to Applications”, PMLR, 2025
    Paper Link | arXiv | BibTeX

  • Increasing Information for Model Predictive Control with Semi-Markov Decision Processes
    R. Hosseinkhan-Boucher, S. Douka, O. Semeraro, L. Mathelin
    Proceedings of the 6th Annual Learning for Dynamics & Control Conference, PMLR, 2024
    Paper Link | arXiv | BibTeX

  • Evidence on the Regularisation Properties of Maximum-Entropy Reinforcement Learning
    R. Hosseinkhan-Boucher, O. Semeraro, L. Mathelin
    Optimization and Learning, Springer Nature Switzerland, 2025
    Paper Link | arXiv | BibTeX

Students

Internships

  • 2026: Salma Ouaissi, ENSIMAG (Grenoble INP – Université Grenoble Alpes), 2nd-year Financial Engineering
    Topic: Constrained Learning for Financial Partial Differential Equations: Exact-Constraint Ansatz
    Role: Co-advisor with C.-A. Lehalle.

  • 2025-26: Joachim Jobard, École Centrale Lyon - KTH Royal Institute of Technology
    Topic: Learning-based Dynamic Programming on Functional Differential Equations
    Role: Co-advisor with L. Mathelin and O. Semeraro.

  • 2025: Eliott Pradeleix, École Polytechnique
    Topic: Learning Functional Differential Equations with Signature-based Encoders
    Led to a publication in the Proceedings of Machine Learning Research (PMLR).
    Role: Co-advisor with A. Shilova, L. Mathelin and O. Semeraro.

  • 2023: Stella Douka, Université Paris-Saclay, M.Sc. in Artificial Intelligence
    Topic: Gaussian Process-based Model Predictive Control with Mutual Information criterion
    Led to a publication in the Proceedings of Machine Learning Research (PMLR).
    Role: Co-advisor with L. Mathelin and O. Semeraro.

Teaching

I have been a teaching assistant for the following courses:

Attendance

Peer Review