Jiyang Tang
Research Scientist at Zoom
Bellevue, Washington, United States
Summary
👤
Senior🎓
Top SchoolJiyang Tang is a research scientist with eight years of experience building production-ready speech and language systems, currently working on contextual biasing at Zoom. He combines academic rigor from Carnegie Mellon and Duke with hands-on engineering experience at startups and industry leaders like Intel and ASAPP, where his work on contextual biasing reduced rare-word ASR errors by up to 49.3% and led to an ICASSP publication. A contributor to the widely used ESPnet end-to-end speech toolkit, he focuses on robust data preparation, tokenization fixes, and reproducible model recipes. He has founded product-focused engineering at Beacon—building a Zoom bot and real-time LLM interview insights—and previously shipped pronunciation coaching apps and APIs for large consumer platforms. Comfortable spanning research, tooling, and deployment, he brings a pragmatic approach to turning speech research into scalable systems.
8 years of coding experience
5 years of employment as a software developer
Bachelor of Science - BS, Data Science, Bachelor of Science - BS, Data Science at Duke Kunshan UniversityBachelor of Science - BS, Data Science, Bachelor of Science - BS, Data Science at Duke UniversityMaster of Science in Intelligent Information Systems, Master of Science in Intelligent Information Systems at Carnegie Mellon University
Chinese, English
Stackoverflow
706reputation
303kreached
19answers
11questions
Github Skills (21)
python10
scripting10
data-preprocessing10
script10
dataprep10
sh10
shell10
data-processing9
speech-recognition8
map-function6
flask6
speech-synthesis6
linux6
forloop6
machine-translation6
Programming languages (21)
Github contributions (5)
espnet/espnet
May 2021 - Jan 2023
End-to-End Speech Processing Toolkit
Role in this project:
Back-end Developer
Back-end DeveloperContributions:33 reviews, 30 commits, 14 PRs in 1 year 7 months
Contributions summary:Jiyang primarily focused on modifying and debugging scripts related to the end-to-end speech processing toolkit. Their contributions include fixing bugs in text tokenization, adding and modifying scripts for data preparation, and fixing typos in shell scripts. They also added scripts for specific data preparation, demonstrating a focus on improving the data handling and processing pipeline for the project.
speech-recognitionspeech-separationchainerspoken-language-understandingspeech-processing
tjysdsg/espnet
May 2021 - Mar 2024
Contributing to ESPnet
Contributions:12 reviews, 35 PRs, 652 pushes in 2 years 9 months
espnet
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.