Run my Keras deep learning model on Google ML Engine and serve predictions.

I am in need of developer who has experience with Python, Keras and Google ML Engine. I have a deep learning model developed in Keras. I want to deploy it on Google ML Engine to serve predictions. I have written the code in a way that it can be deployed on Google ML Engine. The model trains alright when I run it locally using the python script but it has some issues when I use the `gcloud ml-engine train` command. The following are the issues that need to be fixed: 1. The `gcloud ml-engine local train` command to train the model locally runs but fails to install the required dependencies. 2. Currently, the model works by first downloading its training dataset to a local folder from my Google Storage Bucket and training on it. However, this makes a problem with ML engine since it doesn't have a local folder. Ideally it should get data directly on Google Storage bucket instead of downloading it locally but I am open to other suggestions on its solution. 3. I have a local script function that allows you to call predictions on the trained model. I want the serving function implementation relative to Google ML Engine so predictions can be directly from Google ML Engine deployed model. All the documentation, README, and instructions at each step will be provided written in a full guide. I will personally also be available to answer any questions you might have. The budget of this project is $70.