Gesture recognition is a pivotal area of research, especially in the domain of sign language recognition, where precise recognition of gestures is crucial for effective communication. In this study, a novel approach for Indian sign language gesture recognition is introduced, leveraging the shifted window transformers (SWIN) architecture. SWIN Transformers, a recent advancement in computer vision, offer a scalable and efficient solution for capturing spatial dependencies in images. This research work provides a detailed overview of the SWIN architecture and its suitability for gesture recognition tasks, emphasizing its ability to handle long-range dependencies efficiently. A standard ISL dataset is utilized which comprises of diverse hand gestures, with 36 classes comprising of alphabets and digits. Thorough experiments demonstrate the effectiveness of the SWIN-based approach in achieving state-of-the-art results in gesture recognition. This study contributes to the advancement of gesture recognition research, particularly in the context of Indian scenario, by showcasing the capabilities of transformer-based architectures in handling spatial recognition tasks effectively by achieving a classification accuracy of 99.3%