In response to the growing challenges of waste management, this study delves into leveraging advanced technologies to streamline recycling processes. Utilizing the WaRP (Waste Recycling Plant) dataset, available on Kaggle, our research focuses on waste detection and classification within an industrial waste sorting plant context. The dataset encompasses 28 recyclable waste categories, such as plastic and glass bottles, cardboard, detergents, canisters, and cans, presenting complexities like overlap, deformation, and varying lighting conditions. For classification, we utilize the WaRP-C dataset, comprising cut-out image areas with class labels, and experiment with various deep learning architectures, including CNN, Le-Net5, AlexNet, VGG16, Mobile-Net-v2, Inception, and DenseNet, to assess their efficacy in accurate waste classification. In parallel, for waste detection, we employ the YOLO v8 architecture due to its efficiency in cluttered and complex scenes. Notably, synthetic data is incorporated to enhance detection quality, though it results in increased precision for some classes at the expense of decreased recall for others.  The original dataset includes 2452 training images and 522 test images, each annotated with bounding boxes for precise object localization.  Our approach integrates both classification and detection tasks, aiming for a holistic automation of waste sorting processes to advance sustainable waste management practices.