Wen-chin Huang
Assistant Professor at Nagoya University
Nagoya, Japan
Summary
🤩
Rockstar🎓
Top SchoolWen-Chin Huang is an assistant professor at Nagoya University with 11 years of experience in speech technologies, holding an MS and PhD from the same institution and a BS from National Taiwan University. His research and engineering work span voice conversion, speech synthesis, and objective speech quality assessment, with contributions to widely used open-source toolkits such as S3PRL and ESPnet where he implemented any-to-any voice conversion pipelines and production-ready training/evaluation scripts. He has industry research experience from Meta, Facebook Reality Labs, and Google Japan, with multiple papers at INTERSPEECH and ICASSP, reflecting a strong bridge between academic research and real-world systems. Known for both back-end ML engineering and DevOps for speech pipelines, he often focuses on practical model training, custom decoding, and evaluation automation rather than only algorithmic novelty.
11 years of coding experience
3 years of employment as a software developer
Doctor of Philosophy - PhD, Informatics, Doctor of Philosophy - PhD, Informatics at 名古屋大学Bachelor's degree, Computer Science and Information Engineering, Bachelor's degree, Computer Science and Information Engineering at 國立臺灣大學
Chinese, English, Japanese
Github Skills (19)
pytorch10
python10
evaluation10
scripting10
machine-learning10
voice-conversion10
eval10
script10
trainings10
data-processing10
sh10
shell10
modeling10
speech-processing9
self-supervised-learning9
Programming languages (5)
Github contributions (5)
espnet/espnet
Mar 2020 - Nov 2021
End-to-End Speech Processing Toolkit
Role in this project:
Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:154 commits, 9 PRs, 106 comments in 1 year 8 months
Contributions summary:Wen-chin's commits focus on initializing and configuring shell scripts for a speech processing toolkit. These scripts involve setting up environment variables, defining configurations, and preparing data for various speech processing tasks, including feature generation, data preparation, and model training. The user also worked on building the pipelines and preparing the data to be usable by the different models present in the repository.
speech-recognitionspeech-separationchainerspoken-language-understandingspeech-processing
s3prl/s3prl
Aug 2021 - Jan 2022
Self-Supervised Speech Pre-training and Representation Learning Toolkit
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:13 reviews, 13 commits, 1 PR in 5 months
Contributions summary:Wen-chin's contributions primarily involve the development of voice conversion models within the S3PRL framework. Their work includes initializing and modifying model files, specifically related to any-to-any voice conversion tasks using Tacotron2 architecture. The commits focus on adjusting configurations, modifying training scripts, and incorporating custom decoding procedures, indicating a focus on model training, evaluation, and deployment within the project's voice conversion pipeline. The user also makes modifications to the evaluation scripts, suggesting they are involved in the model's performance assessment.
vq-apcspeech-representationdecoarrepresentation-learningspeech-recognition
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.