ML.NET Sample similar to Unity's ML Agents?

I'm looking for a sample I could use to train some game agents/enemies.
So the enemies have the ability to move, dodge roll, melee attack and fireball attack.
During the learning process, 2 enemies are expected to attack each until they learn the possible "win" states.
After a period of learning, I'd then like to take the results and feed them into my game, for the enemies to use.

I essentially want to replicate the ML learning shown in this Unity ML Agents tutorial -
https://www.youtube.com/watch?v=ZZa0faTjwFA

but using ML.NET. Is this possible??
I'm happy to work with someone on this, if needed :D

The issue names no sample, file, or test; start by reviewing the repository's existing ML.NET samples and the linked Unity ML Agents tutorial. Clarify the expected two-enemy training workflow, supported game integration, and what a complete sample would need to demonstrate.

Written by the indexing model from the issue text.