Yin-jyun L
Research Scientist at Spotify
Stony Stratford, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolYin-jyun L is a research scientist at Spotify London specializing in generative AI and representation learning for music and audio, focused on disentangling pitch, timbre and other factors to enable more interpretable and modular audio tools. With eight years of research experience spanning PhD work at QMUL and labs across Asia, he has published at venues including NeurIPS, IJCAI, ISMIR and ICASSP and contributed to real-world music transcription tooling like Omnizart. He has combined academic rigor with industry practice through internships at Stability AI, Sony AI and AIST, developing neural audio codecs and decomposition/recomposition methods for musical mixtures. His background ranges from signal-processing roots in violin mistake detection to modern deep generative models, giving him a rare cross-disciplinary fluency in music, ML and audio engineering. Based in Stony Stratford, he brings both production-minded code contributions (refactoring, checkpointing, deployment-focused changes) and theoretical advances in disentangled representation learning.
7 years of coding experience
6 years of employment as a software developer
Master of Science - MS, Music Informatics, Master of Science - MS, Music Informatics at National Chiao Tung UniversityBachelor of Science - BS, Laser and Optical Technology/Technician, Bachelor of Science - BS, Laser and Optical Technology/Technician at National Central UniversityDoctor of Philosophy - PhD, EECS, Doctor of Philosophy - PhD, EECS at Queen Mary University of London
Github Skills (6)
machine-learning10
python10
modeling9
trainings9
tensorflow9
inference9
Programming languages (3)
Github contributions (5)
Music-and-Culture-Technology-Lab/omnizart
Nov 2020 - Apr 2021
Omniscient Mozart, being able to transcribe everything in the music, including vocal, drum, chord, beat, instruments, and more.
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 7 reviews, 33 commits in 5 months
Contributions summary:Yin-jyun's commits primarily focus on adapting the existing "vocal_frame" module within the "omnizart" project. They modified code to integrate with a new data structure ("MIR1KStructure"), and updated training and inference functions. Furthermore, they addressed merge conflicts, refactored imports and variable names, and incorporated default checkpoints, suggesting a focus on refinement, maintainability, and model deployment within the music transcription project.
everythingmozartbeatpianoable
yjlolo/vae-audio
May 2019 - May 2020
Variational auto-encoders for audio
Contributions:6 commits, 1 PR, 6 pushes in 1 year
source-separationauto-encodersmfccaudiovariational-autoencoder
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.