Kuan Chen
Researcher at ByteDance
Shenzhen, Guangdong Province, China
Summary
🤩
Rockstar🎓
Top SchoolKuan Chen is a speech synthesis researcher and engineer with nine years of experience building production-grade TTS and neural vocoders across major tech companies including ByteDance, Tencent, and Microsoft. He helped develop HiFiNet and lightweight end-to-end models that ran on mobile and Raspberry Pi, and contributed to state-of-the-art open-source TTS work by integrating Tacotron2 and Multi-Band MelGAN for Chinese in the popular TensorFlowTTS project. At Tencent he led speech tech for games (including a Blizzard Challenge second place) and now continues research at ByteDance, blending applied research with product-focused deployment. His background combines an MS in Computer Science from Shanghai Jiao Tong University with hands-on algorithm and system engineering, and he often bridges dataset/ preprocessing work with model inference and deployment — a detail that explains his success shipping both research wins and production systems.
9 years of coding experience
5 years of employment as a software developer
Bachelor, Materials Engineering, Bachelor, Materials Engineering at Huazhong University of Science and TechnologyMaster, Computer Science, Master, Computer Science at Shanghai Jiao Tong University
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (7)
speech-to-text10
text-to-speech10
speech-synthesis10
tensorflow10
taco10
python10
tflite8
Programming languages (7)
Github contributions (5)
TensorSpeech/TensorFlowTTS
Jul 2020 - Aug 2020
:stuck_out_tongue_closed_eyes: TensorFlowTTS: Real-Time State-of-the-art Speech Synthesis for Tensorflow 2 (supported including English, French, Korean, Chinese, German and Easy to adapt for other languages)
Role in this project:
ML Engineer
ML EngineerContributions:22 commits, 1 PR, 15 pushes in 1 month
Contributions summary:Kuan primarily contributed to the implementation of a Chinese text-to-speech example within the TensorFlowTTS framework, focusing on integrating Tacotron2 and Multi-Band MelGAN models. Their work included modifications to preprocessing steps, configuration files, and dataset loading, and encompassed changes to model inference and decoding. They also added support for the Baker dataset.
chinesesynthesisstuckspeech-recognitionstate-of-the-art
azraelkuan/asvspoof2017
Dec 2017 - Jan 2018
Contributions:15 commits, 1 push, 1 comment in 25 days
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.