Tensorflow Machine Learning Project

I need a python script that does the following: - Using tensorflow - Uses CSV files for input and output (example will be provided) Based on arguments to the python script it can do two different things: - Create models for Regression and Classification - Input is a file location to a CSV file with training data - Predict for new data based on the models created - Input is the models to use file location to a CSV with new data to be predicted on (same format as training data Creating Models Ability to create two different types of models - Regression: Predict the goal difference for the game based on the data provided - Classification: Predict if the game will be a WIN / DRAW / LOSS - I would also need the probability of that being true - Also interesting in looking into using neural networks based on your feedback - Need a way to save models and use them for predicting later - Output is the model saved in the file system Prediction - Ability to run the models on new data to predict based on the different types - Output is a new CSV with new prediction columns added. - Ie, PrediectedGoalDiff, PredictedResult, PredictedResultProbability This is a rough design for what I am looking for. Your feedback and suggestion would be appreciated. At the end of the initial version I would like to take over the project but if we both agree it would nice to have you available for help by chat on an hourly basis.