PhD student developing generative AI methods for unsupervised anomaly detection in medical imaging- models that can flag rare or subtle pathologies without relying on the large annotated datasets that are rarely available in real clinical practice. This builds directly on four+ years as a Data Scientist working at the intersection of ML and public health, where I developed and deployed models that shaped real responses to HIV, cancer, tuberculosis, and public health emergencies. That work is where the research question came from: how do you build AI that's genuinely useful when the data isn't perfect, and annotations are scarce? Always happy to connect with people working on AI for health, generative modeling, or anything at that intersection.