Thomas Vuillaume

Astro-physi-data-scientist. Definitely scientist.

Hi. I am Thomas Vuillaume, a data scientist and research software engineer working at LAPP, CNRS.

I hold a master in engineering and a PhD in astrophysics that led me to data science. I develop data analysis pipelines, including machine and deep learning methods, to extract information from the Cherenkov Telescope Array (CTA) currently under construction. I also help researchers and scientists to build better and more open science software and tools.
I am leading or actively participating in several projects to achieve these goals:

I work in a very collaborative environment including a large variety of profiles and often supervise students. Communication and pedagogy have been central in all my professional experiences.
I am passionate about mountaineering sports, photography and travels.

Grants

Teaching

Curriculum rapidum

  • December 2020 - Now : Research Engineer at LAPP, CNRS
  • March 2016 - November 2020 : Postdoctoral position at LAPP, CNRS
  • January 2016 - February 2016 : Postdoctoral position at IPAG, CNRS
  • October 2015 : PhD in Astrophysics, Université Grenoble Alpes
  • 2012 : Master degree Astrophysique, Plasmas, Planètes, Université Joseph Fourier, Grenoble.
  • 2012 : Engineering diploma (Master degree) from Phelma, Grenoble INP. Specialization in Physics and Nanosciences.

Bibliography

ORCID iD icon Google Scholar Badge

My complete list of publications is available on Google Scholar

Portfolio

Here is some of my work or projects I mainly contributed to.

GammaLearn à l'affiche

GammaLearn @IN2P3 ou comment nous utilisons l'intelligence artificielle et la puissance de calcul de Jean Zay pour analyser les images de CTA

Une histoire de Crabe aux Canaries...

crab detection

Un résumé de mes activités fin 2019 aboutissant à la détection de la nébuleuse du Crabe par le LST1, premier télescope du réseau CTA (Cherenkov Telescope Array).

A stratified jet model for AGN emission in the two-flow paradigm

two flow model

Vuillaume et al, 2018. Published in A&A.

GammaLearn

GammaLearn is a framework to apply deep learning techniques to the analysis of low-level Imaging Atmospheric Cherenkov Telescopes such as The Cherenkov Telescope Array The project focus both on improving reconstruction performances but also providing an environment for the analysis of CTA data with DL.

GammaBoard

GammaBoard

GammaBoard is a dashboard developed for Cherenkov Telescopes reconstruction pipelines The dashboard helps keeping track of machine learning experiments and display high-end results in a notebook app to be easily compared.



hipeCTA is a Python 3 library providing High Performance computing algorithms for the Cherenkov Telescope Array (CTA) low-level data analysis. It takes advantage of the latest SIMD (Single input multiple data) operations included in modern processors, for native vectorized optimization of analytical data processing. It is developed to be used under the ctapipe framework.


pschitt! is a Python package for the modelling of atmoSpheric Showers and CHerenkov Imaging Terrestrial Telescopes. pschitt! is a software to image an object in the sky by an array of ground telescopes. It is intended to model atmospheric showers and image them with Imaging Atmospheric Cherenkov Telescopes. "pschitt" is a common French onomatopoeia for the sound produced when a rapid production of gas flows, for example when opening a bottle containing carbonated water.

ctaplot

GammaBoard

ctaplot ctaplot is a collection of functions to make instrument response functions (IRF) and reconstruction quality-checks plots for Imaging Atmospheric Cherenkov Telescopes such as CTA.

Neuron Detector

Neuron Detector provides macros for fiji to analyze videos and images of neurons activations.

Modeling the Emission of Active Galactic Nuclei at Fermi's Era

My PhD thesis realised at IPAG, CNRS between 2013 and 2016 under the supervision of Gilles Henri and Pierre-Olivier Petrucci

Mathematical optimisation based on genetic algorithms.

Mathematical optimisation is the selection of a best element (with regard to some criterion) from some set of available alternatives. The most common way to minimise a function in science is to use the gradient descent method. Unfortunately, this method has its limitations and is not always applicable. I successfully applied genetic algorithms to the research of best models for the modeling of AGN jets emission. See the research section for more information.

Variation of bulk Lorentz factor in AGN jets due to Compton rocket in a complex photon field

Bulk equilibrium Lorentz factor for different external photon fields

Study of the acceleration of AGN jet through Compton rocket. Published and available in A&A Volume 581, September 2015 and on arXiv.

More info on my research

Out of the lab

📷

I am passionate about photography. My work has been presented in several contests, exhibitions, and in national newspapers. It has been awarded several times, including the 1st price at Focales en Vercors, 1st place for winter sport photography and Sports photographer of the year at the International Photo Awards.
You may find it on my website www.thomasvuillaume.com

🏔 🎿 🪂 🧗

Mountaineering helps me escape and fill up the batteries.
I practice regularly hiking, climbing, alpinism, paragliding and ski touring in the Alpes around my home.

Get in touch

Contact me at any time. Robots do not sleep.