Experience

Sicara

• France
• IT Services and IT Consulting
• 1 - 100 Employee

• Lead Data Scientist

  • Jan 2022 - Jul 2023

  Lyon, Auvergne-Rhône-Alpes, France

• Senior Data Scientist

  • Nov 2020 - Jan 2022

• Data Scientist

  • Feb 2019 - Nov 2020

  Région de Paris, France



EDF

• France
• Real Estate
• 100 - 200 Employee

• PhD Candidate in applied mathematics and computational sciences at EDF R&D

  • Mar 2015 - Feb 2019

  Région de Paris, France My research consists in developing model order reduction methods which allows to compute accurate surrogate models of industrial devices. More precisely, I applied these techniques to electrical engineering devices such as transformers and generators in order to speedup 3D nonlinear simulations with high number of degrees of freedom. Finally, these methods enable to construct a library of fast and accurate models which then can be coupled with electrical network softwares such as EMTP-RV… Show more My research consists in developing model order reduction methods which allows to compute accurate surrogate models of industrial devices. More precisely, I applied these techniques to electrical engineering devices such as transformers and generators in order to speedup 3D nonlinear simulations with high number of degrees of freedom. Finally, these methods enable to construct a library of fast and accurate models which then can be coupled with electrical network softwares such as EMTP-RV (http://emtp-software.com/). These methods take advantage of experimental or numerical data in order to reduce the computational complexity of a system. When a large amount of data is available, non intrusive methods such as low rank approximation techniques are particularly adapted. On the contrary, intrusive approaches can be extremely efficient for systems on which a limited quantity of data is provided, especially by capitalizing on the engineer knowledge. Although the aim of my PhD is to apply and adapt these methods to electromagnetic devices, these mathematical approaches are very general and can then be applied to many different domains as soon as some data is available. Show less



Université de Lille Sciences et Technologies

• France
• Higher Education
• 1 - 100 Employee

• Temporary lecturer

  • Sep 2015 - Dec 2017

  Lille I taught computer sciences to undergraduate students (C / Matlab) and electric sciences to undergraduate and graduate students.