Worked on backend reliability and data integrity improvements across the apache/flink-cdc and apache/amoro repositories, focusing on bug fixes that enhanced production robustness. Addressed dependency management in flink-cdc by updating shell scripts to recursively discover all JAR files, ensuring complete library paths during deployment. In apache/amoro, resolved UUID type conversion issues in Iceberg tables by introducing automatic type replacement and a record wrapper, improving data handling for partitioned and bucketed tables. Additionally, improved error messaging for startup mode validation in flink-cdc, clarifying user feedback and reducing setup confusion. Utilized Java, Shell scripting, and data engineering skills throughout these contributions.