[Eridrus]

This is going to be a pretty interesting case study in deploying ML in adversarial contexts :)