Manuel Pariente
AI R&D Manager SmartEyewearLab at EssilorLuxottica
Valbonne, Provence-Alpes-Côte d'Azur, France
Summary
🤩
Rockstar🎓
Top SchoolManuel Pariente is an AI R&D manager and audio researcher with a decade of experience building real-time speech enhancement and source separation systems for hearing applications. He co-founded Pulse Audition to commercialize AI-powered smart glasses that restore selective listening for people with hearing loss, and now leads AI efforts at EssilorLuxottica's SmartEyewearLab. A former Inria PhD student and co-creator/maintainer of the widely used Asteroid PyTorch toolkit, he has contributed advanced filterbanks and separation modules (e.g., DPRNN) that bridge research and deployable systems. His background blends cognitive science, acoustics and deep learning, and he routinely moves algorithms from papers into open-source code and embedded prototypes. Notably, he combines startup grit with academic rigor, mentoring teams that have won competitions like the PyTorch Hackathon while keeping active ties to open-source research.
10 years of coding experience
5 years of employment as a software developer
Master of Science - MS Cognitive Science, Master of Science - MS Cognitive Science at Ecole normale supérieureFirst year Master's degree Applied Physics, First year Master's degree Applied Physics at ENS Paris-SaclayCPGE Mathematics and Physics, CPGE Mathematics and Physics at CPGE Lycée ThiersMaster's degree Acoustics signal processing and computer science, Master's degree Acoustics signal processing and computer science at IRCAM
Stackoverflow
660reputation
70kreached
12answers
1question
keras
top-5%
Github Skills (14)
pytorch10
machine-learning10
signal-processing10
keras9
neural-network6
mean-square-error6
argparse6
merge6
deep-learning6
lstm6
tensorflow6
python6
numpy6
theano6
Programming languages (13)
Github contributions (5)
asteroid-team/asteroid
Oct 2019 - Jun 2022
The PyTorch-based audio source separation toolkit for researchers
Role in this project:
ML Engineer
ML EngineerContributions:23 releases, 318 reviews, 731 commits in 2 years 8 months
Contributions summary:Manuel's primary contribution was to implement new filterbanks and source separation models, demonstrating expertise in audio processing and machine learning. The user added several filterbank implementations, including FreeFB, ParamSincFB, and AnalyticFreeFB, and contributed to the STFT filterbank. Additionally, the user integrated new modules, such as the DPRNN and LSTMMasker, for improved source separation performance, and designed and implemented various changes for model structure. The user's contributions show a focus on research and development within the audio source separation domain.
pytorchsource-separationdeep-learningaudio-separationaudio-source-separation
mpariente/pytorch_stoi
Apr 2020 - Jan 2021
Contributions:1 release, 3 reviews, 50 commits in 8 months
deep-learningpytorchloss
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.