Worked on the xupefei/spark repository to address a critical issue in Apache Spark’s executor memory management when multiple resource profiles are active. Focused on backend development using Scala, the contribution involved fixing a bug that caused incorrect memory configuration, ensuring that executor memory settings reset appropriately based on the currently active resource profile. This targeted change prevents memory allocation errors and enhances the stability of Spark deployments that utilize multiple resource profiles. The solution was implemented directly in the core Spark module, demonstrating a deep understanding of Spark’s resource management internals and careful attention to robust, production-grade system behavior.