Tensorflow federated via gRPC reverse proxy over WebSocket

TensorFlow Federated enables federated machine learning, but uses gRPC for communication. We need to make this data transit through a WebSocket.

The goal of this task is to create a simulated federated learning model, similar to this one: https://github.com/tensorflow/federated/blob/master/docs/tutorials/building_your_own_federated_learning_algorithm.ipynb

Once the model is up and running with two simulated client server and one simulated master server, create the architecture described in the attached figure.

This job is open to a developer who has experience using TensorFlow federated in Python only.