I am a computational biomedicine researcher and Data Scientist focused on building scalable, transparent AI systems for healthcare. My work integrates multimodal patient data, including electronic health records, physiological waveforms, imaging, genomics, and clinical notes, to support earlier detection of critical illness, improve clinical decision-making, and strengthen care delivery across inpatient and outpatient settings.
My research centers on developing robust, interpretable predictive models for conditions such as sepsis, respiratory failure, and cardiometabolic deterioration, with the goal of identifying risk hours before these conditions become life-threatening. A central principle of my work is that clinical AI must be understandable and trustworthy, not a black box. I use methods such as SHAP, attention visualization, and causal feature attribution to make model predictions transparent, help clinicians understand the basis of each recommendation, and support safe, equitable deployment.
