Antoine Liutkus
Algorithm Developer at Hudson River Trading
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolAntoine Liutkus is an algorithm developer and research scientist with a decade of experience blending probabilistic signal processing, machine learning and audio source separation into practical systems. He holds a PhD from Télécom Paris and has moved from industry audio research at Audionamix to academic work at Inria and applied ML roles at Hudson River Trading, where he now develops models for financial signals. Antoine has re-implemented many state-of-the-art audio separation methods—from sinusoidal models and tensor factorizations to deep learning—and has connected source separation to Gaussian process regression and compressed sensing in his research. He contributes to open-source projects like the well-known Open-Unmix PyTorch repository, improving evaluation and robustness for music source separation. Known for digging into why heuristics work, he combines strong theoretical grounding with hands-on engineering across high-dimensional tensor methods and nonparametric stochastic models.
10 years of coding experience
14 years of employment as a software developer
Lycée Henri IVMaster ATIAM, Acoustique Traitement du Signal et Informatique Appliqués à la Musique, Master ATIAM, Acoustique Traitement du Signal et Informatique Appliqués à la Musique at Université Pierre et Marie Curie (Paris VI)Engineer, computer science, signal processing, Engineer, computer science, signal processing at Telecom ParisHabilitation à diriger des recherches, Mathematics and Computer Science, Habilitation à diriger des recherches, Mathematics and Computer Science at University of MontpellierDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Télécom ParisLycée Sainte-Genevièvecollège du jeu de mail montpellierlycée jean monnet montpellier
French, English, Spanish, German
Stackoverflow
1reputation
127reached
1answer
1question
Github Skills (8)
pytorch10
machine-learning10
audio-processing10
python10
testing9
argument-parsing8
keras6
tensorflow6
Programming languages (7)
Github contributions (5)
sigsep/open-unmix-pytorch
Jun 2019 - Oct 2020
Open-Unmix - Music Source Separation for PyTorch
Role in this project:
ML Engineer
ML EngineerContributions:4 reviews, 27 commits, 2 PRs in 1 year 4 months
Contributions summary:Antoine's commits primarily focused on updating and modifying the `test.py` and `eval.py` files within the music source separation project. These updates involved incorporating changes related to the "norbert" module and implementing softmask functionality. The user made improvements to file handling, model loading, and argument parsing for enhanced usability and robustness of the separation process.
pytorchsource-separationmusic-generationopen-unmixmusic-source-separation
aliutkus/beads-presentation
Mar 2018 - Apr 2018
Contributions:32 commits, 4 pushes in 22 days
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.