Amidst escalating data growth, effective classification in diverse domains, including the news industry, is imperative. However, relying solely on human intervention for classification is unfeasible. Addressing the complexities of the Telugu language and leveraging Natural Language Processing (NLP), this study employs classification techniques. Custom Machine Learning and Deep Learning models are developed, utilizing various word embeddings, aiming to enhance accuracy and efficiency in categorizing newspaper articles. The research tackles challenges of unstructured text, attributes, NLP techniques, missing metadata, and algorithm selection. The proposed model offers both generality and efficiency, systematically classifying text documents and demonstrating significant improvements in accuracy through innovative techniques.