For this course we as a team developed SAIS, an application that helps nurses with administrative work by transcribing nurse-patient conversations through automated speech recognition (Whisper V2) and structuring them with the ChatGPT-4 API into a usable interface.

Building SAIS deepened my T&R through working with machine-learning services in Python. But the underlying thoughts was in M,D&C, handling sensitive medical data forced me to weigh model choice as a design decision. Choosing between an API and a local system meant reasoning about where data is stored, who can reach it, and what privacy failure would cost the people involved.