About Me

I am currently Associate Professor at Télécom Saint-Etienne and part of the Data Intelligence Team.
I obtained my PhD in Computer Science from the University Lyon 2 in 2021, after 3 years in the ERIC lab in Lyon, France.
My main research interest are Responsible Machine Learning and Natural Language Processing.

A propos

Je suis actuellement maître de conférences à Télécom Saint-Étienne, et je fais parti de l'équipe Data Intelligence.
J'ai obtenu mon doctorat en informatique de l'université Lumière Lyon 2 en 2021, après 3 ans au laboratoire ERIC à Lyon, France.
Mes sujets de recherche sont l'apprentissage automatique responsable et le traitement automatique du langage.

Publications

Google Scholar

Projects

ANR DIKé : biases and fairness of compressed language models
DaVinciDoc (i-démo project): Agentic corpus exploration, with Wikit and Efalia
ANR Famous : Fairness of multimodal machine learning models
Past projetcs
MOLE (FIL project): Model Optimization via Learning Examples
QABot (regional project) : Chatbot integrating QA technology, with QACompany ant Wikit
ANR Lifranum : Analysis of online litterature

PhD students


Clovis Varangot-Reille : Frugal Routing Strategies for LLMs, with François Jacquenet, Baptiste Jeudy, Christophe Bouvard.
Valentin Rieu : Frugal Alignement of LLMs, with François Jacquenet, Baptiste Jeudy.
Past Students
Dr Thibaud Leteno : Fairness of compressed laguage models, with Christophe Gravier and Charlotte Laclau.
Dr Kunpeng Guo: Domain-Specific Question-Answering, with Christophe Gravier and Dennis Diefenbach.

Post-Doctoral Researcher/Research Engineers

Martin Blanchard : Biases of GNN + LLM models, with Christine Largeron.
Amgad Khalil : Frugal Retriever in RAG pipelines, with François Jacquenet, Baptiste Jeudy, Christophe Bouvard.
Sayan Chaki : Ethic in Multi-Agent settings, with Julien Velcin.
Irina Proskurina : Fairness self amplification by training on synthetic data.
Tristan Caldiere : Biases of LLM based recruitement systems, with Christine Largeron.
Past
Emma Genthon : Biases of PEFT methods
Raphael Chevasson : Reducing Hallucination of LLMs

Come work with me :-)


I may have open positions. Contact me for further informations.

Additional Links


Cordée de la réussite


Optimization - Part 1