Junyi Li
Postdoctoral Research Fellow at City University of Hong Kong
Hong Kong, China
Summary
🤩
Rockstar🎓
Top SchoolJunyi Li is a postdoctoral research fellow with nine years of experience bridging academic research and practical machine learning engineering. Having held postdoc positions at the National University of Singapore and now City University of Hong Kong, Junyi combines deep academic training (doctoral work at Université de Montréal and Renmin University) with hands-on model development. He contributes to open-source text-generation tooling—most notably improving RNN and Transformer decoders and adding Transformer encoders and sequence-to-sequence support in the TextBox library—demonstrating a focus on model architecture and reproducible research. Based in Hong Kong, he blends rigorous experimentation with backend engineering skills, comfortable taking models from prototypes to usable libraries. A detail that stands out: his dual doctoral affiliations suggest a global, interdisciplinary perspective that informs both theoretical and applied NLP work.
9 years of coding experience
Doctor's Degree, Doctor's Degree at Université de MontréalDoctor's Degree, Doctor's Degree at Renmin University of China
Github Skills (10)
text-generation10
transformer10
pytorch10
deep-learning10
python10
seq2seq10
natural-language-processing9
n9
natural-language-generation9
rnn-model9
Programming languages (3)
Github contributions (5)
RUCAIBox/TextBox
Nov 2020 - Mar 2021
TextBox 2.0 is a text generation library with pre-trained language models
Role in this project:
Back-end Developer
Back-end DeveloperContributions:2 releases, 131 commits, 2 PRs in 4 months
Contributions summary:Junyi primarily contributed to the modification and improvement of language models within the project. The commits reveal work on refining components like RNN and Transformer decoders, as well as the introduction of a Transformer encoder. These changes indicate a focus on enhancing the architecture and functionality of the text generation models. Furthermore, the user integrated sequence-to-sequence models, expanding the project's capabilities.
pytorchnlppythonpre-trained-language-modelsnatural-language-generation
turboLJY/CapsGNN-Review-Generation
Oct 2020 - Nov 2020
Contributions:1 release, 31 commits, 14 pushes in 1 month
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.