Portrait
As a passionate engineer and an expert in audio and speech signal processing, I have a lot of experience in sound source separation, localization and HMM based speech recongition.
A pioneer in the application of machine learning methods (NMF) to sound source separation, I am also co-founder of the startup Audionamix (2003).
From 2008 to August 2021, I collaborated with the most prestigious audio signal processing labs in France, such as Télécom ParisTech and IRCAM, notably to apply my skills to the fields of robotics, music, spatialized audio and voice identity conversion.
From September 2021 to April 2022, I worked at the startup Elsa (English Language Speech Assistant) as a Speech Scientist.
Considering a career change to cybersecurity, I am currently working as a Machine Learning Scientist at IRCAM on the production of audiovisual deep-fakes.
Research interests
- machine learning
- Deep Neural Networks, Adversarial Learning
- classification using Support Vector Machines etc
- Non negative Matrix Factorization
- statistical signal processing
- probabilistic modeling
- Bayesian statistics
- audio and speech processing
- speaker localization
- voice identity conversion
- speech and audio source separation