Worked on the apache/spark repository to enhance resource management in Spark SQL by developing robust shuffle file cleanup features. Focused on both adaptive and non-adaptive execution paths, the work ensured consistent removal of shuffle files, reducing unnecessary storage usage and improving deployment safety. Implemented runtime configuration options for shuffle cleanup modes and expanded unit test coverage to validate behavior across various execution scenarios. Introduced an automatic cleanup mechanism for SQL command executions, streamlining resource usage. All changes were delivered using Scala and Spark, with an emphasis on backend development and big data processing, contributing to maintainable and reliable code.