徐亮
Java Developer at Moumou
Beijing, China
Summary
🤩
Rockstar🎓
Top School
Top expert inAdvanced Chinese Natural Language Processing and Machine Learning Technologies
徐亮 is a Java developer based in Beijing with 9 years of professional experience and a strong secondary focus in NLP and deep learning. He contributes to prominent open-source Chinese language projects—such as ALBERT for Chinese and the CLUE benchmark—implementing whole-word masking, Pre-LN adaptations, and distributed training scripts for TPU and PyTorch/TensorFlow. At Moumou he blends backend Java engineering with practical ML tooling, bringing hands-on experience in data pipelines, model pretraining, and text classification systems. His GitHub work shows depth across architectures (TextCNN, BiLSTM, RCNN) and pragmatic data engineering for large-scale language-model training. He pairs an economics background with engineering rigor, which helps him bridge product needs and model evaluation in production contexts. Colleagues would describe him as a pragmatic engineer who quietly advances Chinese NLP infrastructure behind the scenes.
9 years of coding experience
1 year of employment as a software developer
Bachelor's degree, Economics, Bachelor's degree, Economics at Yan bian University
Github Skills (21)
transformers10
pytorch10
python10
pre-trained-model10
machine-learning10
rnn-model10
n10
text-classification10
tensorflow10
bert10
data-processing10
nlp10
benchmark9
multilabel-classification9
tpu9
Programming languages (3)
Github contributions (5)
brightmart/text_classification
May 2017 - Sep 2022
all kinds of text classification models and more with deep learning
Role in this project:
Back-end Developer
Back-end DeveloperContributions:297 commits, 14 PRs, 292 pushes in 5 years 4 months
Contributions summary:徐亮 appears to be working on the core logic and model implementations for text classification. They modified the `p9_BiLstmTextRelationTwoRNN_model.py` file, suggesting a focus on deep learning models involving recurrent neural networks. They also added RCNN implementation and updated TextCNN training scripts, indicating contributions across different model architectures within the repository's text classification framework. Furthermore, the user added multilabel support, improving the usability of TextCNN.
pythonword-embeddingstensorflowtextcnnnatural-language-processing
CLUEbenchmark/CLUE
Nov 2019 - Nov 2022
中文语言理解测评基准 Chinese Language Understanding Evaluation Benchmark: datasets, baselines, pre-trained models, corpus and leaderboard
Role in this project:ML Engineer
ML EngineerContributions:133 commits, 10 PRs, 135 pushes in 3 years
Contributions summary:徐亮's commits primarily focus on modifying and updating scripts related to the creation and execution of pre-training data and fine-tuning tasks for the CLUE benchmark. The user has modified files involving ALBERT model, TensorFlow and PyTorch frameworks, and included scripts for tasks using TPU. The user appears to be focused on integrating and testing pre-trained models in a distributed training environment.
chinesenlubertbenchmarklanguage-understanding
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.