Experience

Ph.D. Student (Conservatoire National des Arts et Metiers)

• January 2023 - Now Paris
• Ph.D. student working on domain adaptation and self-supervised learning for earth observation, as part of the ANR MAGE (Mapping Aerial imagery with Game Engine data) at Conservatoire National des Arts et Métiers. Supervised by Nicolas Audebert (IGN, CNAM), Nicolas Thome (Sorbonne University), Marin Ferecatu (CNAM).

Research Traineeship (Sony CSL)

• April 2022 - September 2022 Paris
• Research traineeship on Diffusion Models applied to expressive musical performances generation. I focused on controllable generation with conditional generation methods. I achieved bibliographical research on the state of the art, implemented various deep learning models and trained them. – Code

Research Project (IRCAM)

• September 2021 - March 2022
• Research project as part of the Centrale Lille research program. We use deep-learning models for music performance modeling. I designed and trained generative adversarial networks (GANs) for this purpose. The sound synthesis relies on Google Magenta’s DDSP model, the core of our problem was then the modeling of the expressive time-series of fundamental frequency and loudness. This work was accompanied by a report and a technical defense submitted to a jury. – Code

Research Traineeship (IRCAM-Pixtunes)

• March 2021 - September 2021 Paris
• Research internship in the ACIDS research team (IRCAM) in connection with the company Pixtunes GmbH. Research and implementation of Machine Learning algorithms for expressive playback of midi files using the DDSP (Differentiable Digital Signal Processing) model. In addition, I did a bibliographic research on the field and prepared two different datasets. During this internship I used the pytorch framework (and pytorch-lightning) and I focused on autoregressive approches and Denoising Diffusion Probabilistic Models (DDPM). – Code

Research Traineeship (INRIA)

• January 2019 - February 2019
• I carried out an internship in a research team in computer sciences named LINKS from Inria laboratories. I was selected to draft a summary of a research paper entitled “Incomplete Data: What Went Wrong, and How to Fix It” written by Leonid Libkin, University of Edinburgh.

Education

Ecole Centrale de Lille (M.Sc.)

• September 2018 - September 2022
• Machine Learning, Data Science, Deep Learning, Signal processing, Acoustics, Computer science, Artificial intelligence, Multi-agent systems, Telecommunication systems, Electronics, Sociology of organisations, Project management.

ATIAM Master Degree (M.Sc.)

• September 2020 - September 2021
• Second year of Master’s Degree in sciences applied to musical applications. Core subjects are : Signal Processing, Acoustics, Machine Learning, Deep Learning, NMF, Data Science for Music, MIR.

CPGE MPSI-MP

• September 2016 - September 2018
• 2-year advanced undergraduate studies in Mathematics, Physics and Computer Sciences. The main purposes of these classes is to develop methods and acquire scientific knowledge and prepare for the nationwide competitive examinations for the entry to the top engineering schools in France.