Problem Statement

Question 2 Context Handling returned merchandise well matters a great deal to any online retail platform. A shopper might ask to send an item back if it fails to meet what they expected. Once a return request comes in, the platform needs to decide what happens to it next. Some requests can go through automatically, some need a human to look them over, and some must be turned down outright. This decision is not random. It is based on: Return eligibility rules, for example, whether the category allows returns and whether the request is within the allowed time window, and Risk signals, such as the customer's historical return behavior across related accounts. Issue The current return routing engine is broken. It sometimes greenlights requests that should have failed a policy check, miscalculates risk scores, routes requests to the wrong outcome, and skips or picks the wrong reason code. Your job is to repair the engine so that it produces the right decision, score, and reason for every re