Developed and integrated a benchmarking environment for the facebookresearch/param repository, focusing on performance profiling and system integration using Python. The work centered on connecting the Param benchmarking tool with Geordi, enabling profiling traces to be uploaded to Manifold or exported locally based on configuration. This integration established an early validation point for Param ahead of the dataswarm pipeline, supporting cross-environment observability and iterative performance assessment. By laying the groundwork for scalable benchmarking workflows, the implementation accelerated data-driven tuning and performance validation. The approach emphasized modular integration and flexible trace management, leveraging benchmarking and profiling skills to enhance future observability capabilities.