Worked on the Azure/azhpc-images repository to deliver a targeted network performance enhancement for high-performance computing workloads. Focused on extending TCP buffer settings for both GB200 and GB300 systems, the work aimed to increase throughput and efficiency in demanding environments. Additionally, NUMA tuning was applied specifically to GB300 to improve memory locality and overall system performance. The implementation was consolidated into a single commit, reflecting a methodical approach to system administration and performance tuning. Utilizing bash scripting as the primary language, the developer addressed HPC image provisioning requirements without introducing bug fixes, demonstrating depth in low-level system optimization and configuration.