At Audax Labs, I developed an auto-tagging system for object detection using YOLOv9. I collected labeled images for classes like guns, bullets, x-rays, and MS COCO data from Roboflow, which streamlined dataset organization and preprocessing. I restructured the dataset to integrate new classes, used Albumentations for image augmentation to handle class imbalances, and split the data into training and test sets. The YOLOv9 model was trained with a custom configuration and evaluated using precision, recall, and accuracy metrics. This system significantly improved the accuracy of detecting objects like weapons in x-ray images, reducing manual effort and enhancing operational efficiency, which strengthened security workflows and supported faster threat detection in critical environments.