In the field of public health, cardiovascular diseases (CVDs) pose a significant burden globally. AI-enabled robust CVD identification, monitoring, and alerting can assist and enhance early screening and hence patient outcomes. In this work, interpretable features are computed from the segments of ECG signal to detect multiple cardiac disorders in patients. A segment is a time window that contains number of beats. It makes the segment-level analysis more robust in comparison to a beat-level study. The morphological characteristics of the individual beat are also extracted from the ECG signal, and irregular heartbeats are identified to help medical professionals gain a detailed understanding. In this work benchmark public datasets of congestive heart failure, sleep apnea, arrhythmia, and normal sinus rhythm is used. The performance of the disorder detection is quite high for both segment-level and beat-level analysis. The proposed system’s lightweight design (non-deep ML) and clinically interpretable features improve the chance of future clinical translation due to enhanced transparency, and therefore improved trustworthiness.