Developed a configurable ANTLR DFA cache size feature for the Spark SQL parser in the apache/spark repository, focusing on improving memory management and parsing stability for large or complex SQL workloads. The solution introduced a clear configuration interface, allowing production clusters to tune DFA cache limits without code changes, thereby reducing the risk of out-of-memory errors during SQL parsing. Leveraging Scala, SQL, and software testing skills, the work enabled more predictable memory usage and improved throughput under peak loads. The feature was delivered end-to-end, preparing Spark deployments to better handle memory pressure events during intensive parsing operations.