The project develops a wearable solution for context-aware detection of cardiac anomalies based on ECG signals acquired via a BLE-enabled device and processed on a smartphone using lightweight neural networks and Federated Learning. The system combines continuous ECG monitoring with smartphone inertial data and, when necessary, heart rate estimation via PPG to improve detection robustness. The project outcome is a functional TRL 45 prototype of a wearable system with embedded artificial intelligence.