On a global level, driving in a drowsy state is responsible for many Road accidents. To further enhance this, we employ CNN i.e. Convolutional Neural Network to develop a real-time system for detecting driver drowsiness. This system assists to track a driver through a webcam, significantly concentrating on the face as well as the closing of eyes and yawning to ascertain to what extent the driver is alert. A pre-trained CNN model with OpenCV support for face landmarks assists to classify states of “alert” and “drowsy” effectively. When drowsy, an alert system with audio-visual signals is initiated. Furthermore, responses to external hardware like buzzers for better application are achieved by Arduino. This work aptly focuses on the importance of real-time tracking, dependability, and embedded systems with deep learning for more effective execution of the driver security system with a feasible cost.