Jules SintesPhD Candidate in Reinforcement Learning (MARL)

I am a PhD candidate in the ARGO Team at INRIA Paris and École Normale Supérieure (DIENS, PSL University) under the supervision of Prof. Ana Bušić. My research focuses on multi-agent reinforcement learning (MARL) on network structured problems with an emphasis on energy systems applications. I am particularly interested in theoretical aspects of decentralized MARL in fully cooperative problems. I aim at understanding how one can leverage underlying structures of particular problems to learn joint optimal policies with emerging cooperation behavior. I also work on MARL with delays, a very common problem in real systems such as wind farms.

Industrial Experiences

Prior to the beginning of my PhD, I worked during 2 years as a Data-Scientist in the industry. I was involved in various industrial AI projects:

  • Reinforcement Learning and Deep Learning for topology management recommender system for powergrid in close collaboration with RTE (French Transmission System Operator).
  • Automatic Speech Recognition and Speaker Diarization for the Senate (upper house of the French Parliament).
  • Search engine for information retrieval in large document corpus and Retrieval Augmented Generation for Cartier.
  • Various small internal R&D projects around Natural Language Processing, Large Language Models and Computer Vision.

Before that, I worked during one year as a R&D and production engineer at Syos, a startup that creates tailormade 3D printed saxophone mouthpieces.

Open-Source Contributions

As an Open-Source enthusiast, I also contributed to some projects and published open-source packages, datasets and models.

Music

Besides research, I play the saxophone in several bands and compose music.