Yongqiang Wang
Principal Research Scientist at NVIDIA
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolYongqiang Wang is a Principal Research Scientist with over a decade of experience building large-scale deep learning and speech systems across top labs including NVIDIA, Google, Apple, and Meta. He has led efforts to scale ASR and universal speech recognition, delivered world-record training speeds on distributed platforms, and helped build foundation models for multimodal applications. A hands-on researcher-engineer, he contributes to influential open-source projects like TensorFlow Lingvo, extending core sequence-to-sequence and attention components used in speech and translation research. His work combines rigorous academic foundations (PhD, Cambridge) with practical engineering that ships at product scale, and he’s known for turning big data into robust, deployable models that advance cross-lingual communication.
10 years of coding experience
12 years of employment as a software developer
Ph.D, information engineering, Ph.D, information engineering at University of CambridgeBEng., Electronic engineering and information science, BEng., Electronic engineering and information science at University of Science and Technology of ChinaThe University of Hong Kong (HKU)
English, Chinese
Github Skills (7)
machine-learning10
machine-translation10
speech-recognition10
nlp10
tensorflow10
python10
distributed-computing9
Programming languages (2)
Github contributions (5)
tensorflow/lingvo
Oct 2021 - Nov 2022
Lingvo
Role in this project:
ML Engineer
ML EngineerContributions:11 commits in 1 year 1 month
Contributions summary:Yongqiang primarily contributes to improving the Lingvo framework, which is focused on machine learning tasks like speech recognition and machine translation. Their work includes enhancing the `inspect_model` utility to analyze model parameters and tensor sizes, which is critical for model debugging and optimization. The user also implements and refactors core components, such as `ExtractBlockContextV2` and `ChunkwiseAttention` to support new attention mechanisms in the framework. These contributions indicate a strong focus on expanding the capabilities of the Lingvo platform for sequence-to-sequence models.
asrtranslationctcspeech-recognitiontensorflow
yqwangustc/dotfiles
Dec 2016 - Jul 2024
Contributions:15 pushes, 2 branches in 7 years 7 months
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.