Yist Lin
Research Engineer (Computer Vision) at TikTok
Singapore
Summary
🤩
Rockstar🎓
Top SchoolYist Lin is a research engineer specializing in speech processing, NLP, and computer vision with 11 years of experience building production-ready ML systems. Currently at TikTok, he develops multimodal models for content moderation and investigates LLM integration for large-scale training and difficult-case mining. Previously he led breakthroughs in Automatic Speech Translation and ASR—creating semi-supervised pipelines across 40 languages, inventing a non-peaky CTC for better timestamps, and delivering industry-leading auto-dubbing quality. An active open-source contributor to the widely used s3prl toolkit, he brings hands-on model architecture and dataset engineering expertise. With an MS from National Taiwan University and a history of managing GPU clusters and applied research, he bridges research rigor and production reliability.
11 years of coding experience
5 years of employment as a software developer
Master of Science - MS, Computer Science, Master of Science - MS, Computer Science at National Taiwan UniversityInformatics, Informatics at Technical University of Munich
Github Skills (11)
machine-learning10
pytorch10
self-supervised-learning10
modeling9
model-driven-development9
model-driven9
datasets9
speech-recognition9
model-building9
lstm8
python7
Programming languages (4)
Github contributions (5)
s3prl/s3prl
Jul 2020 - Mar 2021
Self-Supervised Speech Pre-training and Representation Learning Toolkit
Role in this project:
ML Engineer
ML EngineerContributions:44 commits, 2 PRs, 24 pushes in 8 months
Contributions summary:Yist primarily contributes to the self-supervised speech pre-training and representation learning toolkit, s3prl. Their commits involve adding and rewriting datasets like Google Speech Commands and QUESST 2014, and refactoring the existing code. The user also modified the model architecture, embedding loss, and evaluation methods, demonstrating expertise in speech processing and machine learning model development.
vq-apcspeech-representationdecoarrepresentation-learningspeech-recognition
yistLin/beamertheme-yeast
Nov 2019 - Dec 2020
Contributions:9 commits, 10 pushes, 9 branches in 1 year 1 month
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.