AI for Healthcare

My research focuses on developing AI methods for learning meaningful patient representations from multimodal biomedical data to advance precision medicine. By integrating longitudinal electronic health records (EHRs), multi-omics, and social determinants of health, I develop deep learning, graph-based, transformer-based, and large language model approaches to capture the complex and heterogeneous nature of human disease. These representations provide a foundation for identifying clinically and biologically meaningful disease subtypes, characterizing disease heterogeneity and progression, and enabling downstream applications such as risk prediction, prognosis, treatment response modeling, and patient stratification.

My long-term goal is to develop robust, interpretable, and transferable AI models that transform multimodal biomedical data into actionable knowledge, enabling more individualized approaches to disease prevention, diagnosis, and treatment.