Drowsy driving is a major concern, responsible for a frightening number of road accidents. This research tackles this problem head-on by proposing a drowsiness detection system (DDS) that leverages the power of machine learning to proactively enhance road safety. This research holds significant value for two key reasons. Firstly, it brings attention to the dangers of drowsy driving and the critical need for preventative measures. Secondly, the proposed DDS offers a extremely precise and instantaneous method to detect drowsiness. By highlighting the role of fatigue in accidents, the research emphasizes the importance of developing solutions to address this specific threat. The system utilizes the convolutional neural network (CNN), specifically the MobileNet architecture. This choice enables the DDS to analyze driver behavior in real-time with exceptional accuracy, reaching a staggering 99.8% success rate in identifying drowsy drivers. This paves the way for practical implementation in driver assistance systems, ultimately contributing to a significant reduction in drowsy driving accidents and a safer driving experience for everyone.