About me

I am currently an "Attaché Temporaire d'Enseignement et de Recherche" at Université Paris Cité at IRIF Lab.

Previously, I was a Postdoctoral researcher at Sorbonne Université on Jussieu campus, under the supervision of Nicolas Perrin-Gilbert, since 2022.

I have been a PhD student at SUMO under the supervision of Thierry Jéron, Nicolas Markey and David Mentré from 2018 to 2022.
My research focuses on formal methods for real-time systems.
I defended my thesis on March 11 2022 (manuscript, defence slides).

News

  • (24/03) Paper submitted at RAMICS 2024
  • (18/03) Enzo Erlich joins Jeremy and I to work on HDA
  • (mid-march) Our paper about HDTA was accepted to Petri Nets 2024 !
  • Research

    My research focuses on: A brief (outdated) summary of what I do can be found here (last update: 13/10/2021) and here is my CV (last update: 29/03/2023).

    Publications

    We can find my publication list on DBLP.

    Implementation

    I have the following projects:
  • Inria gitlab page with timed automata implementation projects written in Python
  • Gitlab page with a robotics implementation project written mostly in Python
  • A grobner basis implementation (current) project, written in Ocaml
  • Students

  • Enzo Erlich : Master 2 Student March 2024 - August 2024, Language of Higher Dimensional Automata (co-supervision With Jérémy Ledent)
  • Talks

    Scheduled talks:

    Previous talks:

    Research projects (Outdated, last update: early 2023)

    Symbolic computation of the maximal permissiveness of timed automata

    This is one of my two phd project, under the supervision of Thierry Jéron, Nicolas Markey and David Mentré. In this project, we model the perturbation of delays in Timed Automata to compute how much we can be permissive during a run. We model it with a two-player game, where a first player proposes an action and an interval of delay (instead of a delay in the classic semantics) and the opponent chooses the delays taht will really apply.

    Our aim is to compute, given an initial location, the optimal strategy of the player to maximize the minimum of the size of all the intervals he have to propose during a run to reach a goal location. We proposed two algorithms to compute this strategy for all valuations, with a permissiveness function that depend on the valuations.

    The first algorithm compute the permissiveness function and the second one the binary and levelled permissiveness functions, that respectivelly computes: I implemented theses two algorithms in python, based on PPLPY, for linear timed Automata.

    This implementation is maintained in this gitlab page Py-Symb-robustness. It was previously maintained in MERCE-FRA (github). Both are under GPL licence.

    Numerical computation of the maximal permissiveness of timed automata

    This is my second phd project, under the supervision of Thierry Jéron, Nicolas Markey and David Mentré. This project consists in computing the permissiveness with a double backtrack algorithm that samples the set of possible intervals and possible delays.

    I have proposed an algorithm, detailled in the 6th chapter of my PHD thesis.

    This algorithm is implemented in python, based on networkx library. It was first maintened on the github of MERCE ( MERCE-FRA (github)). It is now maintained on my gitlab page (Py-Num-robustness).

    Both implementation are under GPL Licence.

    Efficient implementation of control strategy for hybrid systems

    In my post-doctoral project, in collaboration with Nicolas Perrin-Gilbert and Philipp Schlehuber, our aim is to compute the optimal strategy to reach a goal, for systems of timed automata using some stopwatches. We use parameters to optimize our goal by minimizing the global time elapsing during the run and use an SMT-based method to make our model more realistic for our application: robotic collision avoidance.

    This project is implemented in python (not public yet, coming soon after submission).

    Teaching

    At the end of this academic year, I will have done 296 hours of teaching in several topics and to a diverse audience.
    At Paris Cité Université, UFR d'informatique.
    • 2023 -- 2024 (2nd semester): Introduction to Programming 2, practial session & seminars (TD+TP)
      • 48 hours
      • Class: L1 Info.
      • Supervisor: Gulio Manzonetto
    • 2023 -- 2024 (1rst semester): Algorithmics seminars (TD)
      • 24 hours
      • Class: L3 INFO3.
      • Supervisor: Matej Stehlik
    • 2023 -- 2024 (1rst semester): Java seminars (TD)
      • 48 hours
      • Class: two classes: L2 Info2/Bio-Info + L2 Math-Info
      • Supervisor: Cristina Sirangelo.
    • 2023 -- 2024 (1rst semester): Automata and Syntaxic Analysis, seminars & practical sessions (TD+TP)
      • 36 hours
      • Class: L2 (Maths-Info)
      • Supervisor: Ralf Treinen.
    • 2023 -- 2024 (1rst semester): Fête de La Science, Scientific Mediation
      • 10 hours
      • Class: Elementary
    At Rennes:
    • 2019 -- 2021: Java practical sessions ("TP"), computer science bachelor students. 17 hours per year
      • Class: L1 Info, Rennes.
      • Supervisor: Thomas Genet.
    • 2018 – 2020: Java practical sessions and lectures ("CM" and "TD"), bio-informatics master students. 25 hours per year.
    • 2019 – 2020: Mathematics seminars and lectures ("CM" and "TD"), computer science master students. 10 hours
      • Class: M1 Info, ENS Rennes.
      • Co-supervised (Lilian Besson, Joshua Peigner, Bastien Thomas, Emily Clement)
    • 2019 – 2020: Formal language seminars ("TD"), computer science bachelor students. 18 hours
      • Class: L3 MIAGE, ISTIC (Rennes).
      • Supervisor: Mickael Foursov.
    • 2018 – 2019: Formal calculus, seminars ("TD"), computer science bachelor students. 8 hours
    • 2018 – 2019: Python/Processing practical sessions, junior high school students. 15 hours

    Agenda