Abstract: Athenahealth is a leading provider of cloud-based services for medical practices and health systems, with solutions including revenue cycle management, patient engagement and population health management, and electronic medical records (EMRs). This talk illustrates how we use machine learning and athenahealth’s vast network of clinical data to classify medical documents . Doctors using athena Clinicals EMR must tag imaging results with specific labels from an existing compendium, such as ‘x-ray, ankle, view 2.’ To alleviate this repetitive, administrative burden, we deploy ensemble models built using NLP techniques to automate labeling. Preliminary results show that we can identify the correct label from thousands of candidates with high accuracy. We also present progress toward implementing the training and prediction pipeline in a production EMR.