https://github.com/roman-shaban Building control layers for AI systems | Execution, Compliance, Safety