Worked on the google/cel-java repository to deliver a targeted performance optimization for the CelUnknownSet union operation. Refactored the set union logic to leverage Sets.union and SetView.immutableCopy with a pre-sized builder, which reduced internal hash table churn and improved both memory efficiency and CPU utilization. This change enhanced the throughput of set-union operations, directly supporting business goals around faster query processing and reduced garbage collection pressure in high-load, client-facing workloads. The work demonstrated a strong command of Java, performance optimization techniques, and set operations, with careful attention to low-level data structure choices to improve scalability and operational efficiency.