The paper presents a novel research endeavor, unveiling the “Deep Learning based Dynamic Sign Language Recognition System.” This work addresses the intricate challenges inherent in the recognition of dynamic sign language gestures through the lens of advanced deep learning methodologies. Leveraging the potency of convolutional neural networks (CNNs) and recurrent neural networks (RNNs), the system embodies an amalgamation of spatial and temporal features, engendering a heightened discernment of nuanced sign gestures. By harnessing the power of long short-term memory (LSTM) units and attention mechanisms, the architecture achieves proficient sequential modeling and selective feature focus. This technology enables seamless communication for the hearing impaired, enhancing accessibility and fostering inclusive. Users can express themselves naturally, improving their overall quality of life. Furthermore, the study delves into data set curation, preprocessing, and augmentation techniques tailored to the unique attributes of sign language data.