Worked on enhancing error handling and reliability for blob lifecycle operations in the linkedin/ambry repository, focusing on backend development with Java and unit testing. Addressed a critical bug by refining how the system reports and manages permanently deleted blobs, ensuring that undelete requests return HTTP 410 and that purged-from-disk states are correctly mapped to HTTP 404. Added comprehensive regression tests to validate delete and undelete flows, clarifying HTTP semantics and reducing ambiguity for clients. These improvements help maintain data consistency, minimize client confusion, and support more effective incident response, ultimately contributing to a more robust and maintainable backend system.