This internship report describes work done at CEAT Ltd. to improve the accuracy of examining prints on tires using data analytics. Currently, print examination is done manually before mass production, which can be improved. The student worked on a text extraction model using deep learning to detect and extract printed text from tire images. Over 4000 tire images were manually annotated using label tool software to create a training dataset. The model aims to identify regions of interest in images and extract printed text to automate and improve the accuracy of print examination.