CV

Here is a brief summary of my professional experience. You can download the full PDF on the top right of this page.

Contact Information

Name Dimitri Lerévérend
Professional Title Ph.D. Student in Privacy-Preserving Decentralized Learning
Email dimitri.lereverend@inria.fr

Professional Summary

I am a Ph.D. student in privacy-preserving decentralized learning at INRIA Rennes. My research focuses on privacy techniques for decentralized learning, with a particular emphasis on noise correlation. I work on both theory and practice to design approaches that are scalable, tractable, and supported by strong privacy guarantees.

Experience

  • 2023 - present

    Rennes, France

    Ph.D. Student
    INRIA Rennes
    Supervisors: Davide Frey, Romaric Gaudel, François Taïani
    • Conducting research on decentralized learning privacy and collaborating with researchers across multiple institutions.
  • 2025 - 2025

    Lausanne, Switzerland

    Research Intern
    Sacs Lab, EPFL
    Decentralized learning convergence under fragmentation. Supervisor: Anne-Marie Kermarrec
    • Studied theoretical convergence rates of decentralized learning with model fragmentation.
  • 2023 - 2023

    Rennes, France

    Research Intern
    Team WIDE, INRIA Rennes
    Privacy protection through private averaging. Supervisors: Davide Frey, François Taïani, Romaric Gaudel
    • Studied privacy properties of decentralized machine learning and the effects of private averaging using PyTorch.
  • 2021 - 2021

    Rennes, France

    Research Intern
    D7, IRISA Rennes
    Adapting Public Announcement Logic update mechanisms to Dynamic Epistemic Logic. Supervisor: Guillaume Aucher
  • 2020 - 2021

    Rennes, France

    Research Project
    Percept, INRIA Rennes
    Feature extraction for high dynamic range images. Supervisors: Olivier Le Meur, Mathieu Chambe
    • Built a neural network for HDR image feature extraction in TensorFlow.
  • 2020 - 2020

    Bordeaux, France

    Research Intern
    Mnemosyne, INRIA Bordeaux
    Visualization of echo state networks. Supervisor: Nicolas Rougier
    • Developed a spatialized ESN and visualized activity using Voronoi diagrams.

Education

  • 2023 - present

    Rennes, France

    Ph.D. in Computer Science
    INRIA Rennes
    • Advisors: Davide Frey, Romaric Gaudel, François Taïani
    • Exploring privacy in decentralized learning via noise correlation, from theoretical foundations to practical methods.
  • 2022 - 2023

    Rennes, France

    Master's degree (M2 SIF) in Computer Science
    ENS Rennes
    • Focus on machine learning, image representation, and complex systems.
  • 2021 - 2022

    Rennes, France

    Preparation for the "Agrégation d'informatique"
    ENS Rennes
    • Focus on teaching and computer science fundamentals: algorithms, architecture, operating systems, networking, and databases.
  • 2019 - 2020

    Rennes, France

    Bachelor's degree in Computer Science
    École Normale Supérieure de Rennes

Publications

Certificates

  • Master's degree - University of Rennes (2023)
  • Agrégation d'informatique (Ranked 2nd) - French national education (2023)

Teaching

2024–2026: PRI, M1, Cyberschool: Course presenting privacy formally.
2024–2026: Agrégation preparation, M1, ENS Rennes: Course preparing students for the "Agrégation d'informatique", the national exam for teaching computer science in France.
2025–2026: ISE, L1, Cyberschool: Basic introduction to security with a mini-project.
2023–2025: ACE, L1, ISTIC: Course explaining algorithm complexity practically.
2024–2025: Trustworthy AI, M1, Master IA: Course introducing trustworthy AI: privacy and explainability.
2023–2024: BDD, L2, ISTIC: Course presenting databases.

Presentations

2026: ICLR: Poster Presentation, Rio
2026: Transnumériques: Oral Presentation, Rennes
2026: YSP: Oral Presentation, Paris
2025: EurIPS: Workshop Presentation, Copenhagen

Skills

Programming Languages: Python, OCaml, C, C++, Java
Libraries: scikit-learn, TensorFlow, PyTorch, pandas, NumPy

Languages

French : Native speaker
English : Fluent

References

  • Upon request

    Email me if you require a list of references.