During this period, work centered on enhancing the apache/incubator-gluten repository by implementing configurable compression for the VeloxColumnarBatchSerializer. The solution introduced a Spark configuration parameter, allowing users to select the desired compression codec, with corresponding updates in the C++ backend to apply the chosen algorithm. Comprehensive test suites were developed in Scala and C++ to validate behavior across all supported compression methods, ensuring correctness and stability. This feature targeted memory optimization for operations such as broadcast joins, enabling per-operation compression choices and improving scalability. The implementation adhered to repository conventions and included CI-friendly tests for maintainability and future validation.