Worked on enhancing the apache/gravitino repository by implementing end-to-end persistence of Paimon bucket distribution metadata through the Flink connector. Developed a solution in Java that enables bucket-key and bucket options specified in Flink SQL to be stored and round-tripped via Gravitino metadata, ensuring accurate reconstruction through SHOW CREATE TABLE. The approach involved introducing distribution parsing, relaxing validation rules, and extending both catalog and test coverage to support scenarios like AUTO and bucket-only distributions. Demonstrated backend development and data engineering skills by integrating Flink, SQL, and metadata modeling to improve reliability and usability for distributed data workloads.