Northwestern Medicine built a transformer model in Epic that flags radiology follow-ups, tracked with Weights & Biases.

Only about two-thirds of imaging follow-up recommendations are addressed, risking delayed diagnoses. Northwestern built a Results Management AI that finds follow-up findings in radiology reports and notifies physicians through Epic within three minutes, using W&B for experiment tracking and GPU monitoring. The page reports nearly 5,000 physician interactions and more than 2,400 follow-ups tracked to completion.

Weights & Biases is an MLOps platform for experiment tracking and model monitoring, used by biotech and hospital AI teams building protein and clinical models

This sits in the Ship stage of the AI use-case lifecycle: Named examples from the vendors pharma teams use to get an AI system live inside a real, regulated production workflow — validation, CRM, safety and clinical systems of record.