Wonjin Yoon
Postdoctoral Research Fellow at Harvard Medical School
Boston, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolWonjin Yoon is a Postdoctoral Research Fellow based in Boston with a decade of experience in NLP, BioNLP, and bioinformatics, currently affiliated with Harvard Medical School and Boston Children's Hospital. He holds a PhD in Computer Science from Korea University and has a strong track record of applied research collaborations with industry, including AstraZeneca and NAVER Clova AI. His open-source contributions include significant engineering improvements to the widely used BioBERT project, notably enhancing NER robustness through fixes to tokenization, padding, and detokenization. He combines rigorous academic training with production-minded ML engineering, shipping reproducible pipelines and model improvements tied to peer-reviewed work (e.g., BioBERT, BioCreative/DrugProt). He selectively engages with peers in his research domain, prioritizing collaborations that advance clinical and biomedical text mining. Colleagues value him for bridging cutting-edge language models with practical clinical NLP applications.
10 years of coding experience
Doctor of Philosophy - PhD, Computer Science, 4.4 / 4.5, Doctor of Philosophy - PhD, Computer Science, 4.4 / 4.5 at Korea UniversityExchange student, Computer Science, Exchange student, Computer Science at National University of Singapore
English, Korean
Github Skills (7)
machine-learning10
nlp10
tensorflow10
python10
bert10
natural-language-processing10
bioinformatics9
Programming languages (5)
Github contributions (5)
dmis-lab/biobert
Jan 2019 - Jul 2022
Bioinformatics'2020: BioBERT: a pre-trained biomedical language representation model for biomedical text mining
Role in this project:
ML Engineer
ML EngineerContributions:27 commits, 4 PRs, 22 pushes in 3 years 5 months
Contributions summary:Wonjin primarily contributed to the development and improvement of the NER (Named Entity Recognition) task within the BioBERT project. Their work included fixing padding errors, addressing tokenization issues, and adding features for logit output. The user also updated the NER pipeline, refactored code, and functionized detokenization processes. These changes collectively enhanced the accuracy and functionality of the NER model.
biomedical-text-mining
dmis-lab/bioasq-biobert
Jul 2019 - Jul 2022
Pre-trained Language Model for Biomedical Question Answering
Contributions:18 commits, 4 PRs, 17 pushes in 2 years 11 months
language-modelquestion-answering
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.