Jaewoo Kang

Also known as 강재우, Jae-Woo Kang, Jae-woo Kang

Jaewoo Kang is a computer scientist whose Korea University lab builds language models for biomedicine, from BioBERT to the Meerkat medical models. He earned a Ph.D. at the University of Wisconsin–Madison in 2003, founded the startup WISEngine in 2000, and taught at North Carolina State University from 2003 to 2006. In 2019, he coauthored BioBERT, and his team won BioASQ Task 7B and earned a Best Performer award in the IDG-DREAM drug-kinase binding challenge. His lab introduced Meerkat-7B in April 2024, the first 7B-parameter model to pass the USMLE threshold. He is a professor of computer science at Korea University, supervises its DMIS Lab, and is CEO of AIGEN Sciences.

Now at: Korea University School of Medicine and AIGEN Sciences Inc.
Badges: Founder
Bridge
Boomerang

Timeline

1994–2003

Ph.D. at Wisconsin–Madison

He earns a B.S. in computer science at Korea University in 1994, an M.S. at the University of Colorado at Boulder in 1996, and a Ph.D. in computer science at the University of Wisconsin–Madison in 2003. From 1996 to 1997, he is a member of technical staff at AT&T Labs Research in Florham Park, New Jersey.
2000–2001

Founding WISEngine

In 2000, while a doctoral student at Wisconsin–Madison, he founds WISEngine Inc., a startup in Seoul and Santa Clara that builds Internet data-management technology, and serves as its CTO until 2001. He is one of ten US recipients of a Microsoft Graduate Fellowship in 2000.
2003–2006

Professor at NC State

He joins North Carolina State University as an assistant professor of computer science in 2003 and stays until 2006.
2019

BioBERT for biomedical text

In January 2019, he is a coauthor of BioBERT, a language representation model that adapts BERT to biomedical text by pre-training on large-scale biomedical corpora. BioBERT improves on earlier state-of-the-art models in biomedical named entity recognition, relation extraction and question answering, and Bioinformatics publishes the paper in 2020.
2019

BioASQ and DREAM wins

In 2019, his Korea University team uses a BioBERT-based question-answering model to take first place in all five batches of BioASQ Task 7B, ahead of Google. His team also earns a Best Performer award in the IDG-DREAM Drug-Kinase Binding Prediction Challenge in 2019.
April 2024

Meerkat-7B passes the USMLE

In April 2024, his lab introduces Meerkat-7B, the first 7B-parameter medical language model to pass the USMLE threshold of 60%, with 74.3% on MedQA. The Meerkat models range from 7 to 70 billion parameters, learn from chain-of-thought reasoning paths drawn from 18 medical textbooks, and in the 70B version outperform GPT-4 by an average of 1.3%.
2025

BioASQ 13B and CURE-Bench

In 2025, his DMIS Lab team wins the BioASQ 13B challenge with first place in three of four batches. The team also earns Excellence Awards in three of four categories at CURE-Bench at NeurIPS 2025, a competition on agentic AI reasoning for therapeutic decision-making.
As of October 2026

Professor at Korea University

He is a professor of computer science at Korea University, jointly appointed with the Department of Medicine in the KU School of Medicine, and supervises the DMIS Lab. He is also Chief Executive Officer of AIGEN Sciences Inc. In August 2026, the lab announces four papers accepted to EMNLP 2026, including Meerkat2 on efficient medical AI.