Léo Hemamou

Léo Hemamou

Ph.D. , Reasearch Scientist

Introduction

Hi everyone :)

I am Léo Hemamou, and welcome to my webpage!

I am currently an Applied Scientist at Asteria, a bio-inspired innovation startup. I am focused on bridging the biological and engineering worlds via NMT, KG, etc

In 2023, I was part of the AIDA team (Artificial Intelligence and deep analytics) in Sanofi as an NLP research data scientist.

In 2022, I held a research data scientist position at ICIMS in the Talent Cloud Artificial Intelligence Team.

In 2021, I obtained my PhD in Computer Science entitled “Automatic Analysis of Multimodal Behaviors in Job Asynchronous Video Interview”. My PhD was in collaboration between LIMSI-CNRS, Telecom ParisTech, and EASYRECRUE.

Research topics

Since my thesis I have always been interested in a large number of topics in the field of machine learning. Nevertheless, over the years, my expertise has been sharpened especially on the use of deep learning for the understanding of human behavior and NLP. This field of research touches many topics such as automatic speech analysis, automatic speech recognition, multimodal fusion, interpretability, fairness, long sequence processing, automatic summarization, PEFT, etc

Recently, I am interested in Knowledge Graphs, Small LM and Neural Machine Translation.

PhD Research topics

My PhD research work was pluridisciplinary as it combined IO psychology, machine learning, multimodal model, and social signal processing. I was especially interested in an automatic evaluation of hirability in the asynchronous job interview. Among the topics studied in the context of my Ph.D., I was focusing on personal selection, job interview, neural networks, multimodal representation, and interpretability.

The aim of my work was to :

  1. evaluate automatically the hirability of a candidate based on his video answers
  2. infer social competencies of the candidate
  3. compare the results obtained through machine learning with results in IO psychology, especially for non-verbal behavior.

I’m also interested in Fairness especially disentanglement models and adversarial methods.

Collaborators

My PhD work was supervised by

  • Jean-Claude MARTIN, Computer Science Professor at University Paris-Sud, conducting his research at LIMSI-CNRS.
  • Chloé CLAVEL, Associate Professor in Affective Computing, conducting his research at Telecom ParisTech.

I have intensely worked in collaboration with