Printed Circuit Boards (PCBs) are the essential building blocks of modern electronic devices, providing a critical foundation for signal transmission and electrical connections. However, their complex designs make them prone to defects such as short circuits, spurs, and spurious copper, which reduce device reliability and performance. This paper proposes a new mobile-based deep learning model for detecting defects in PCBs. The model utilizes You Only Look Once version 8 (YOLO8) and a heatmap for visualization. The defect detection model produces annotated images and heatmaps to identify defects in PCBs. Integrated into the mobile application named Qualitronix, it allows technicians to use smartphone cameras for on-site inspections. The application displays heatmaps highlighting concerns and annotated images detailing defect types and locations. By combining computer vision with mobile technology, Qualitronix modernizes PCB quality assurance, providing a scalable and accessible solution. The proposed defect detection model has been evaluated against three benchmark datasets, demonstrating accurate and efficient defect detection. It consistently performs well across various PCB defect detection benchmarks, with precision, recall, and F1-score values ranging from 84.5% to 98.6% on test sets, indicating strong accuracy and generalization capabilities. The Mean Average Precision at an Intersection over Union (IoU) of 0.5 (mAP50) ranges from 85.7% to 98.6%, while the stricter mAP50-95 varies between 50.6% and 75.9%. These results demonstrate the model’s effectiveness in identifying fine-grained defects across diverse evaluation thresholds.