Xiang Liang
ENGINEER&RESEARCHER at CreditEase
Haidian District, Beijing, China
Summary
👤
Senior🎓
Top SchoolXiang Liang is a seasoned engineer and researcher with 14 years of experience specializing in machine learning, data mining, recommender systems, and social network analysis. Currently pursuing a Ph.D. at the Institute of Automation, Chinese Academy of Sciences, he has published work on temporal recommendation models (KDD2010, WI2009) and contributed to production-scale ML at companies including 今日头条, Hulu, and CreditEase. He brings strong C++ skills and practical systems experience—evidenced by significant contributions integrating Baidu's WarpCTC into the widely used Apache MXNet to improve sequence-modeling and speech-recognition workflows. A former member of The Ensemble in the Netflix Prize, Xiang combines deep research pedigree with hands-on engineering to bridge novel algorithms and real-world, high-performance systems.
14 years of coding experience
2 years of employment as a software developer
Bachelor of Engineering (BEng), 自动化, Bachelor of Engineering (BEng), 自动化 at University of Science and Technology of Chinaph.D., Machine Learning and Data Mining, ph.D., Machine Learning and Data Mining at Institute of Automation, Chinese Academy of Sciences
English
Github Skills (12)
ocr10
mxnet10
machine-learning10
deep-learning10
ocra10
python9
gpu9
cuda9
c-language8
rnn-model8
n8
cprogramming-language8
Programming languages (6)
Github contributions (5)
apache/mxnet
May 2016 - Dec 2016
Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more
Role in this project:
ML Engineer
ML EngineerContributions:8 commits, 8 PRs, 129 comments in 7 months
Contributions summary:Xiang contributed significantly to integrating Baidu's WarpCTC library, demonstrating a focus on optimizing the repository's capabilities in the area of sequence modeling and/or speech recognition. Their work included implementing WarpCTC, addressing label size discrepancies, and ensuring compatibility across CPU and GPU environments. This involved code modifications in C++ and Python to integrate the library, adjust configurations, and address potential memory management issues. The user also developed and refined example implementations using WarpCTC, highlighting their focus on practical application and performance optimization.
pythonschedulerdataflowmutationdata-science
xlvector/gogo
Mar 2016 - Apr 2016
Contributions:366 commits, 361 pushes, 2 branches 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.