Summary
Jeonghyeon Kim is a computational biology researcher and graduate student at Duke University with nine years of experience bridging AI, structural biology, and drug design. He develops and adapts state-of-the-art models—modifying AlphaFold2, integrating diffusion models, and combining Graph Neural Networks with reinforcement learning—to generate synthesizable lead compounds and better model protein complexes like pMHC-TCR tricomplexes. His work spans large-scale virtual screening (over one billion molecules) and active learning strategies that cut docking compute by orders of magnitude, alongside multi-task models for kinase off-target prediction. Based in Durham, NC, he brings a rare mix of electrical/software engineering training and chemistry/biological engineering expertise that enables both algorithmic innovation and practical screening pipelines. He contributed methods used in CASP16 and focuses on applying AI across small molecules, antibodies, and immunotherapies to make drug discovery more predictive and efficient.
9 years of coding experience
Master's degree, Chemistry, Master's degree, Chemistry at 서울대학교 (Seoul National University)
Doctor of Philosophy - PhD, Bioengineering and Biomedical Engineering, Doctor of Philosophy - PhD, Bioengineering and Biomedical Engineering at Duke University
English, Chinese