Developed a JDBC-backed storage backend for partition statistics in the apache/gravitino repository, enabling seamless integration with relational databases such as MySQL, PostgreSQL, and H2. The solution provided transactional guarantees for updates, enhancing data durability and supporting robust backup and recovery workflows. By delivering a pluggable configuration, the implementation complemented the existing Lance storage without introducing breaking changes, increasing flexibility for users managing partition statistics. The work demonstrated strong skills in Java, SQL, and database management, with a focus on maintainable software development practices and collaborative code delivery through a merged pull request and co-authored commit in the project.