Zehui Lin
Natural Language Processing Engineer
Shanghai, Shanghai, China
Summary
👤
Senior🎓
Top SchoolZehui Lin is a Natural Language Processing engineer with 10 years of experience, currently leading novel machine translation efforts at ByteDance AI Lab in Shanghai. He blends research and product impact—publishing at ACL/EMNLP and achieving first place in four WMT constrained directions (en→de, fr→de, en→km, en→ps). His engineering work includes contributing a Transformer-based seq2seq decoder to the fastNLP framework, reflecting deep practical knowledge of attention mechanisms and production-ready model components. Trained at Fudan (MS) and Xiamen University (BS), he moves seamlessly between experimental research and deployable systems. Colleagues rely on him to turn state-of-the-art ideas into competitive MT systems and reusable open-source tooling. He is also active in competitive MT benchmarks, which informs his pragmatic approach to model design and evaluation.
10 years of coding experience
1 year of employment as a software developer
Master's degree, Computer Science, Master's degree, Computer Science at Fudan UniversityBachelor's degree, Computer Software Engineering, Bachelor's degree, Computer Software Engineering at Xiamen University
Github Skills (10)
attention-mechanism10
transformers10
transformer10
pytorch10
nlp10
deep-learning10
nlpjs10
python10
natural-language-processing10
text-processing9
Programming languages (6)
Github contributions (5)
fastnlp/fastNLP
Mar 2020 - May 2020
fastNLP: A Modularized and Extensible NLP Framework. Currently still in incubation.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:5 commits, 1 PR in 1 month
Contributions summary:Zehui's commits primarily focus on the implementation of a Transformer-based sequence-to-sequence decoder within the `fastnlp` framework. They introduce new modules, including `TransformerSeq2SeqDecoder`, `TransformerSeq2SeqDecoderLayer`, and related components like `DecoderMultiheadAttention` and `TransformerPast`, to facilitate sequence generation tasks. The contributions demonstrate an understanding of attention mechanisms and transformer architectures, adding capabilities to the NLP framework. The user integrated various components related to sequence-to-sequence modeling.
nlpbertword-embeddingsdeep-learningtext-processing
linzehui/linzehui.github.io
Jul 2018 - May 2021
Contributions:99 commits, 106 pushes, 1 branch in 2 years 10 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.