Worked on improving the reliability of the intel/lkvs performance benchmarking workflow by addressing a critical issue in the benchmarking script. Focused on debugging and scripting using Python and Shell, the developer corrected a typo in the test script to ensure the intended benchmark executed properly. Enhanced error handling by replacing return-based exit codes with sys.exit(1), allowing runtest.py to accurately capture and report failures to continuous integration systems. These changes stabilized the benchmarking process, improved measurement reliability, and increased CI failure visibility. The work demonstrated attention to detail in performance testing and contributed to more robust automated testing pipelines for the repository.