In this project, we trained a Recurrent Neural Network which decodes the original message using a standard handwritten digit classification dataset.The task is to classify a given image of a handwritten digit into one of 10 classes representing integer values from 0 to 9, inclusively. The promise of adding state to neural networks is that they will be able to explicitly learn and exploit context in sequence prediction problems, such as problems with an order or temporal component. Natural Language Processing is one of the core fields for the Recurrent Neural Network applications due to its sheer practicality.The large chunk of business intelligence from the internet is presented in natural language form and because of that RNN are widely used in various text analytics applications.We build and verified our prediction models using Keras on Tensorflow. We were able to achieve 95 % accuracy on test data. We further accelerated the model inference by utilizing the Open Compute Language(OpenCL) which can run on heterogeneous platforms.