During November 2025, Hubert Nadolny contributed to the intel/compute-benchmarks repository by developing an optimized random write kernel for memory benchmarking. He refactored existing tests to eliminate special handling for random buffer contents, introducing a dedicated write_random kernel that streamlined the codebase and improved throughput potential. This work focused on simplifying the benchmark pipeline, enhancing maintainability, and laying groundwork for future kernel-level optimizations. Hubert applied his expertise in C++ development, memory management, and benchmarking methodologies, using disciplined Git-based collaboration. While no major bugs were addressed, his targeted code cleanup supported the optimization and reduced complexity for ongoing and future development.