<p>Cardiovascular Disease (CVD) is one of the major causes of death in the world today, where millions of deaths are caused annually, and this underlines the necessity of early and accurate diagnosis technique.Electrocardiogram (ECG) is a non-invasive method of recording bio-potentials, which is used to record myocardial signals by using surface electrodes.The ECG helps in the early identification of abnormalities in the electrophysiology that are important in minimizing morbidity and mortality associated with CVD. The most important characteristic of this type of analysis is the accurate determination of R-peaks in the QRS complex, which can be used to extract valuable biomarkers such as Heart Rate Variability (HRV). The analysis of HRV comprises both linear (time and frequency domain) and non-linear methods (entropy measures, fractal analysis) dynamics, which help in understanding the intricate relationship between autonomic regulation and heart function. With the advent of computational intelligence, machine learning (ML) and deep learning (DL) have revolutionized the methods of non-invasive data-driven diagnostics in the field of cardiology. In this review, the state-of-the-art ML and DL models (supervised classifiers, Convolutional Neural Networks (CNNs), and Long Short-Term Memory networks) are explored to detect the QRS complex and extract the HRV features and classify cardiac events. It also discusses on several critical issues including signal variation during acquisition, model interpretability, domain generalization, and clinical use which are critical towards bridging the gap between algorithm advancements and useful diagnostic assistance tools in the real world.</p>