In T283526: Create generic revscoring inference service, we created a base image which can serving revscoring models via KFServing.

This means we should be able to move the ORES models over to Lift-Wing, however, due to the large number of models, it makes sense to migrate each class of models so we can watch for edge cases.

The next step is to identify the steps required to migrate the edit quality models. Here is a rough outline of how it might go:

1. Move the model binary files to storage
2. Create a custom resource definition for each model's inference service
3. Figure out CI/CD deployment pipeline