UC Davis Computer Science and UC Davis Health are working to empower people with T1D who inject insulin manually (i.e., do not use insulin pumps) to use a closed-loop style of monitoring and adjustments. They are developing a “metabolic watchdog” system which monitors an individual’s glucose levels, predicts if they are likely to go out of range, and provides them with a behavioral nudge so that they can correct their glucose trajectory. By providing a closed-loop style of management to people injecting insulin manually, they open up new therapy mechanisms that can enable them to have tighter control over their glucose levels without them needing to check their CGMs every five minutes and decide if/when to take action — the metabolic watchdog software will check it for them automatically and provide suggestions when needed. Their core software for predicting glycemic imbalance uses machine learning, trained on data provided by Tidepool. UC Davis and UC Davis Health received a CITRIS Core Seed Funding Grant to fund this project, and their work is supported by the Tidepool Big Data Donation Project.