During December 2024, contributed a performance-focused enhancement to the xupefei/spark repository by developing a read-only configuration access interface for SparkConf. This feature allowed configuration parameters to be accessed without duplicating the entire configuration object, directly addressing latency and memory pressure in Spark’s backend. The work demonstrated a strong command of Apache Spark internals, Scala, and Java/JVM optimization, with careful attention to API design for backend development. By reducing CPU usage and garbage collection overhead, the solution enabled higher throughput for latency-sensitive Spark workloads, reflecting a deep understanding of both system performance requirements and robust software engineering practices.