This document proposes a non-invasive method to detect hypoglycemic events in patients with type 1 diabetes using wearable sensors. The method involves collecting physiological and activity data from sensors like ECG, accelerometers and breathing monitors. Machine learning models analyze the data to detect glycemic events, which are then represented semantically and used to generate alerts. Preliminary tests show the physiological model can accurately detect hypoglycemic events based on continuous glucose monitor data. The system aims to help patients and practitioners monitor insulin levels without invasive blood glucose testing.