Added a hands-on self-containted MLflow/Ray Serve deployment example

[dmatrix]

Signed-off-by: Jules S.Damji [email protected]

Why are these changes needed?

To augment examples in Ray Serve's Integration with Model Registries section,
I added a hands-on self-contained example of how to use mlflow.sklearn.autolog(), which logs
all the metrics, parameters, and artifacts to the local model registry; and a deployment class that
loads the registered model as a pyfunc model and does the prediction.

This complete example augments the previous partial code snippet.

Checks

• [x ] I've run scripts/format.sh to lint the changes in this PR.

[architkulkarni]

Looks great, thanks for adding this! I'll make a followup issue to run this code sample in CI using Sphinx's literalinclude. I'll need to figure out where to install mlflow in the CI setup but it should be straightforward.