Printed circuit boards (PCBs) becoming more complex as technology advances, adding new components and changing their architecture. One of the most crucial quality control procedures is PCB surface inspection since even little flaws in a signal trace may have a significant detrimental effect on the system. It has always been difficult to determine the pass/fail criteria in traditional machine vision systems based on small failure samples, despite the advancements in sensor technology. Suggesting a sophisticated PCB inspection method built on a skip-connected convolutional auto encoder to address these issues suggested to enhance the PCB inspection system by using convolutional autoencoders. The original, fault-free photos and the damaged ones were used to train the deep autoencoder model. The defect location was then located by comparing the decoded images with the input image. Using proper image augmentation to enhance the model training performance in order to get over the tiny and uneven dataset in the early phases of production. Printed circuit boards, or PCBs, are essential parts of electronic gadgets and are very significant to the electronics sector. While ensuring PCB quality and reliability is crucial, manual inspection techniques are often labour and error-intensive. The proposed novel machine learning (ML)-based method for identifying PCB defects demonstrates a significant improvement in detection rates compared to traditional methods, offering a promising solution for the electronics manufacturing industry.