Salah Zaiem
Senior Research Scientist at Google DeepMind
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolSalah Zaiem is a Senior Research Scientist based in Paris with nine years of experience at the intersection of speech and audio-visual machine learning. He progressed from PhD work and internships at Inria, Google Research, and Mila to research roles at DeepMind, where he now leads audio-visual generation efforts. His contributions to the SpeechBrain open-source toolkit—implementing downsampling strategies, CTC support, and efficiency fixes—show practical expertise in PyTorch-based ASR pipelines and model optimization. Trained at École Polytechnique and ENS Paris-Saclay in machine learning and data science, he combines strong theoretical foundations with hands-on engineering. Outside core research he has applied technical and communication skills in diverse settings, from building chatbot platforms to managing large student events, reflecting an ability to translate research into real-world impact. Colleagues would note his blend of rigorous research, production-oriented contributions, and a knack for improving model efficiency that isn’t obvious from publications alone.
9 years of coding experience
4 years of employment as a software developer
Ingénieur Polytechnicien Majoring in Data Science, Ingénieur Polytechnicien Majoring in Data Science at École PolytechniqueTunisian Baccalaureate, Tunisian Baccalaureate at Lycée pilote de l'ArianaPreparatory classes for the highly competitive entrance in French Engineering Schools, Preparatory classes for the highly competitive entrance in French Engineering Schools at Lycée Pierre de FermatMaster MVA Machine Learning, Master MVA Machine Learning at École Normale Supérieure Paris-Saclay
Arabic, French, English, Spanish
Github Skills (9)
pytorch10
speech-recognition10
audio-processing10
deep-learning10
speechbrain10
asr10
neural-network9
machine-learning9
convolutional-neural-networks9
Programming languages (2)
Github contributions (5)
speechbrain/speechbrain
Oct 2021 - Feb 2024
A PyTorch-based Speech Toolkit
Role in this project:
ML Engineer
ML EngineerContributions:4 reviews, 7 PRs, 7 comments in 2 years 4 months
Contributions summary:Salah's primary contribution involved adding and refining downsampling code for an Automatic Speech Recognition (ASR) system built on PyTorch. They implemented various downsampling techniques, including convolutional and signal-based approaches, which suggests efforts to optimize the input audio processing pipeline. Further commits focused on correcting bugs, particularly related to the computation of MACs, which is often used for model efficiency analysis, as well as correcting indentation and flake8 errors. The user also added CTC code for the ASR and made additional minor improvements.
voice-recognitionasrspeech-recognitionspeech-separationspeaker-verification
salah-zaiem/ModalWeb
Mar 2017 - Dec 2018
Contributions:35 pushes, 2 branches in 1 year 9 months
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.