Worked on the apache/spark repository to enhance the robustness of Spark SQL by addressing two critical bugs affecting query reliability under complex schemas. Focused on backend development using Scala and SQL, the developer resolved a hashCode contract violation in BatchScanExec, ensuring consistent behavior in hash-based collections by aligning hashCode with equals. Additionally, they improved the safety of filter remapping after nested schema pruning, preventing FIELD_NOT_FOUND exceptions by securely handling projection failures. Both fixes were validated with targeted unit tests, increasing test coverage and stability. No new features were introduced, but these changes improved Spark SQL’s correctness and runtime predictability.