Modify Stefan's pipeline to use MLflow + TensorFlow model

[piotrpdev]

Stefan's models use sklearn which isn't directly supported by OpenVINO. See this issue discussion for more reasons we are switching.

This is the new model that uses TensorFlow.
That one was too old, here is the actual new model.

User Story

• As a MLOps engineer I would like to take a trained model and use an automated pipeline to package it inside a container so that it can be deployed to the near edge. #2

A/C

• Stefan's pipeline fully works using the new model
• Test data is provided, and the tests work.