Worked on the apache/spark repository to enhance Spark’s Kubernetes integration by developing an Executor Failure Tracker for managing executor pod creation failures. Using Scala and leveraging backend development skills, implemented a mechanism that monitors pod creation attempts and throttles retries after a configurable maximum, thereby reducing unnecessary resource consumption and log noise in constrained environments. The approach included adding comprehensive unit tests to ensure reliability and regression safety, with no user-facing changes introduced. This work improved the reliability and operational efficiency of Spark on Kubernetes, demonstrating a strong focus on test-driven development and robust resource governance within distributed systems.