Professional Experience
PhD in Astrophysics at Institut d'Astrophysique Spatiale, Université Paris-Saclay
Simulation of galaxies using a VAE, conditionned to physical parameters using a regressive flow (MADE). Implementation in the Euclid pipeline. Adaptation of a Probabilistic U-Net for the probabilistic segmentation of overlapping astronomical objects.
Data analysis of a high dimensional dataset. Development of an interactive tool for data visualisation and reproducibility.
Supervised by Marc Huertas-Company, Hervé Dole and Alexandre Boucaud, with long-term stays at IAC (Spain) and APC (France).
2017-2019
Three months internship at Institut d'Astrophysique de Paris, Sorbonne Université.
Calibration of the charge transfer inefficiency for Euclid using Deep Learning algorithms (WGAN).
Supervised by Henry J. McCracken and Tom Charnock.
2019
Two months internship at LERMA, Observatoire de Paris.
Study of the potential correlation between gas dynamics and stellar morphology for galaxies using random forests and convolutional neural networks (classifiers).
Supervised by Marc Huertas-Company.
2018
Various internships (less than 3 months each) at IRAP (Toulouse), APC (Paris) and MPQ (Paris)
Verification of the universe accelerated expansion with supernovae.
Study of the scanning strategy for LiteBird.
Test of a mono atomic-layer machine for supra-conductive studies
2015-2017
Education
Double Master's Degree in Astrophysics and Astronomy, at Observatoire de Paris.
Specialisation in Theory and Data Analysis for cosmology, plasma and galaxies.
2017-2019
Bachelor Degree in Fundamental Physics, at Université de Paris.
Fundamental Physics and Engineering.
2013-2017
Mentoring
Engineering School student (6 months).
Main supervisor. Adptation of our VAE to produce multi-band simulations of galaxies.
2021
Astrophysics Master’s Degree student (6 months)
Co-supervisor. Use of VAE to cluster galaxy spectra in the latent space.
2020
Astrophysics Master’s Degree student (3 months)
Main supervisor. Understanding and use of Self Organizing Maps for galaxy catalogues comparison.
2020
Computer Science Skills
Python
Five-year experience with common scientific libraries (especially numpy, scipy, astropy and pandas) as well as typical machine learning frameworks (scikit-learn, keras, tensorflow).High interest in innovative ways for data visualisation and exploration (matplotlib animations, streamlit, ipywidgets)
Strong expertise on generative and probabilistic models for image processing, built over several interships and PhD work. Adaptation of U-Nets, VAEs, Regressive Flows and SOMs forastrophysics projects. Development of GANs and WGANs from scratch in pure tensorflow
ML/DL
Familiarity with bash and bash scripting. Work on large clusters with job scheduling (with slurm).
Fortran90, C, C++: classes during Master’s degree, used for short academic projects.
Languages
Daily basis: Jupyter, Visual Studio Code, LATEX, Blender, Keynote.
Astrophysics: TOPCAT, SExtractor, Galapagos, Galsim.
Softwares
macOS on laptop, Ubuntu, CentOS on servers.
OS
Everyday use of Git for code development. Collaborative workflows on GitHub and GitLab.
Versionning
created with
Website Builder Software .