Worked on expanding hardware support and improving test reliability in the pytorch/benchmark repository. Developed HPU device integration within the Benchmark Testing Framework, enabling users to run benchmarks on HPU devices through a new CLI option and ensuring automatic device recognition. Addressed test reporting accuracy by updating the test runner to use pytest.skip for skipped tests, which improved metadata and categorization in test results. Leveraged Python, Pytest, and CI/CD practices to deliver these enhancements. The work enabled broader, hardware-agnostic benchmarking and provided clearer, more reliable test outcomes, supporting maintainable test automation and more robust device support within the project.