Research

Optics · Space Instrumentation · Machine Learning · Image Processing

DISPERS — PSF Modeling for Euclid NISP (PhD project, 2022-2025)

The DISPERS project at CPPM is dedicated to the characterization and modeling of the instrumental response of the NISP (Near Infrared Spectrometer and Photometer) instrument aboard the Euclid space telescope (ESA, launched July 2023).

Euclid will map the large-scale structure of the universe over more than 15,000 square degrees of sky to probe dark energy and dark matter. Accurate characterization of the instrument's Point Spread Function (PSF) is critical for the scientific exploitation of its data.

My contributions within DISPERS:


Computer Vision & Deep Learning

Beyond my main PhD research, I explore computer vision and deep learning applied to various image analysis tasks: object detection & segmentation (YOLO, Faster R-CNN), image classification via transfer learning, and generative models (VAEs, autoencoders). I also apply deep learning to astronomical image analysis — light curve analysis and period detection for exoplanet transits from Kepler/K2 and TESS data.


Wine Aging Prediction with MLPs

A personal side project applying multi-layer perceptrons (MLPs) to predict wine quality and aging potential from physicochemical properties (acidity, alcohol, pH, sulphates, etc.). This project sharpened my skills in data preprocessing, hyperparameter tuning, feature engineering, and model interpretability (SHAP values).


Astronomy — IPSA VEGA Association

As Vice-President of the IPSA VEGA astronomy association, I led data reduction pipelines for Kepler/K2 and TESS photometric data: aperture photometry, light curve detrending, Lomb-Scargle periodograms, and asteroseismic analysis. I also authored articles for astronomy magazines and taught astrophotography and astrophysics courses to association members.