Design an ML system that detects harmful content at scale, with a focus on weapon-sales posts, harmful marketplace listings, and harmful ads. The system should classify user-generated content or ads, decide whether the content violates policy, and route the item to the appropriate downstream action.
