Fix my simple RNN deep learning model and make it exportable via tensor-flow CLI

I have written a simple RNN model in tensorflow using the 'High Level API' architecture style so that it can be used on Google ML Engine. I have a working version of the same model written in Low Level API architecture but I was unable to write it properly in Hight Level API style. My model is an exact RNN replica of the this CNN: [login to view URL] The model learns on patterns of randomly generated numbers and then is able to detect right pattern from wrong pattern after training. I want the developer to fix the model and also make it exportable via saved_model_cli of tensorflow so I can use it on Google ML Engine. Also add the documentation to Readme file on how to use it. The budget of this project is $70. Developer should be previously experienced with Tensorflow along with the coding standards.