Xinghai Sun
Expert Researcher at Tencent
Beijing, China
Summary
🤩
Rockstar🎓
Top SchoolXinghai Sun is an expert researcher and team leader based in Beijing with nine years of experience building large-scale ML systems for recommendation, search, speech, and ads. Currently at Tencent he leads research and engineering for video recommendation and has driven RL research in Tencent AI Lab, after building distributed deep learning infrastructure and production ML for Baidu Research. He combines hands-on model work—contributions to PaddlePaddle’s PaddleSpeech and official models, including DeepSpeech2 data pipelines and a memory-augmented machine translation prototype—with product-facing system design. His background spans multimedia search and computer vision from internships and research at Microsoft Research Asia and NUS, plus an MS in Computer Science from Peking University. Known for translating research into deployable pipelines, he blends algorithmic depth with practical engineering for high-throughput services. An understated strength is his track record of improving training-to-inference workflows, from SortaGrad optimization to inference tooling in open-source speech toolkits.
9 years of coding experience
4 years of employment as a software developer
Bachelor's degree, Automation Engineer Technology/Technician, Bachelor's degree, Automation Engineer Technology/Technician at University of Electronic Science and TechnologyMaster's degree, Computer Science, Master's degree, Computer Science at Peking University
Github Skills (16)
neural-network10
paddlepaddle10
machine-translation10
machine-learning10
speech-recognition10
nlp10
deep-learning10
trainings10
python10
asr10
modeling10
data-pipelines9
data-pipeline9
n9
rnn-model9
Programming languages (4)
Github contributions (5)
PaddlePaddle/models
May 2017 - Nov 2017
Officially maintained, supported by PaddlePaddle, including CV, NLP, Speech, Rec, TS, big models and so on.
Role in this project:
ML Engineer
ML EngineerContributions:143 commits, 86 PRs, 42 pushes in 6 months
Contributions summary:Xinghai contributed model configuration and code for a machine translation model with external memory within the PaddlePaddle models repository. Their work involved creating and modifying Python files, specifically `mt_with_external_memory/mt_with_external_memory.py`, to implement the model's architecture and components. The commits demonstrate an understanding of neural network architectures for sequence-to-sequence tasks, including the use of external memory mechanisms for enhancing model performance.
nlpdeep-learningpaddlepaddlerecommendationcomputer-vision
PaddlePaddle/PaddleSpeech
May 2017 - Nov 2017
Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text frontend, Speaker Verification System, End-to-End Speech Translation and Keyword Spotting. Won NAACL2022 Best Demo Award.
Role in this project:Backend Developer & ML Engineer
Backend Developer & ML EngineerContributions:107 commits, 1 push, 2 branches in 5 months
Contributions summary:Xinghai contributed to the development of the DeepSpeech2 model within the paddlespeech repository. The commits show the addition of a LibriSpeech dataset and an audio data provider, alongside the simplification of a DeepSpeech2 model configuration, suggesting the user worked on the model's data loading pipeline. Further commits included the implementation of inference functionality and the application of SortaGrad optimization, which highlights the user's focus on model training and deployment.
voice-recognitionasrtranslationspeech-recognitionstreaming-tts
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.