Hojae M. Choi

I have three years of industry experience as an AI research engineer, focusing on developing AI systems for drug discovery. In addition, I’ve spent over five years conducting research on machine learning and statistical models in the biomedical field.

My work spans from developing deep learning and machine learning models to MLOps—training and evaluating AI/ML models—as well as designing and building web applications to support their deployment and use. I’ve collaborated closely with medicinal chemists and CADD engineers both within and outside the company, contributing to data-centric strategies and the design and analysis of experiments.

I’m particularly interested in representation learning for proteins and small molecules, multi-objective optimization, and uncertainty estimation. Currently, my research focuses on generative models for drug design, especially on tackling practical challenges to enhance their usability in real-world applications.

During my time at KAIST, I developed machine learning models for gene prioritization and diagnostic biomarker discovery using graph neural networks, integrating genomics and metabolomics data through biological networks.

I enjoy working at the intersection of science and technology, and I’m always eager to learn, share ideas, and contribute to meaningful research. I primarily work with Python and am experienced with various AI/ML libraries in cloud-based environments.