Doctoral student Junjie Luo was first author of a paper published in the Nature Partner Journal: Health Systems (“A Large Sensor Foundation Model Pretrained on Continuous Glucose Monitor Data for Diabetes Management“). This article presents a new approach to glucose prediction for diabetes management that harnesses the pretraining technique exemplified in large language models. The authors hypothesized that a large sensor model pretrained on continuous glucose monitor data could learn glucose fluctuation patterns and predict glucose values. Using a large dataset with 15.96 million glucose records from 592 diabetic patients, Luo and colleagues expanded the use of AI for diabetes management from long-term risk level prediction to near future prediction. The team used JHU’s Rockfish and Discovery high-performance computing (HPC) clusters as computational resources in this study.