During February 2026, Dzeri focused on enhancing storage efficiency and reliability for Spark on Kubernetes within the apache/spark repository. He developed a feature enabling the reuse of Persistent Volume Claims (PVCs) with equal or greater capacity, supporting dynamic volume expansion and reducing over-provisioning. This approach preserved shuffle data during auto-expansion, improving resilience and lowering cloud storage costs. Dzeri implemented the solution in Scala, leveraging his expertise in Kubernetes and cloud computing, and ensured robustness by adding comprehensive unit tests to validate the new logic. His work demonstrated depth in Kubernetes PV management and contributed measurable business value through cost efficiency.