Chin-yun Yu
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolChin-yun Yu is a doctoral student in computer science at Queen Mary University of London with a decade of engineering and research experience in audio signal processing and machine learning. They have combined industry roles at HTC, Rayark, and a research internship at SonyAI with academic work on autoregressive neural networks for music and multi-pitch estimation at Academia Sinica. As an ML engineer contributor to torchaudio, Chin-yun optimized the lfilter implementation—adding conv1d-based acceleration, batching, and improved backprop support—demonstrating a knack for turning research insights into performant open-source code. Comfortable across back-end systems and deep learning models, they fuse practical production experience with rigorous research, while moonlighting as a bedroom musician and guitarist who brings hands-on audio intuition to technical problems.
10 years of coding experience
4 years of employment as a software developer
Bachelor's degree, Computer Science, GPA 3.87/4.3, Bachelor's degree, Computer Science, GPA 3.87/4.3 at National Chiao Tung UniversityDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Queen Mary University of London
Chinese, English
Github Skills (6)
audio10
pytorch10
machine-learning10
audio-processing10
python10
backpropagation9
Programming languages (8)
Github contributions (5)
pytorch/audio
Feb 2021 - Aug 2021
Data manipulation and transformation for audio signal processing, powered by PyTorch
Role in this project:
ML Engineer
ML EngineerContributions:41 reviews, 16 commits, 26 PRs in 6 months
Contributions summary:Chin-yun primarily focused on improving the `lfilter` function within the `torchaudio` library. Their contributions involved replacing operations with more efficient ones like `conv1d`, enhancing backpropagation support, and adding batching and filterbank support. The user's work indicates efforts to optimize the filtering process and extend its capabilities within the audio processing domain.
audiopytorchsignal-processingpythonio
yoyolicoris/torch-fftconv
Mar 2021 - Dec 2021
Implementation of 1D, 2D, and 3D FFT convolutions in PyTorch. Much faster than direct convolutions for large kernel sizes.
Contributions:4 releases, 1 review, 61 commits in 9 months
3dfftkernelpytorch
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.