During January 2026, Keyur upgraded the Kineto submodule within the pytorch/pytorch repository to the latest commit, focusing on enhancing profiling reliability and runtime stability. He utilized C++ development skills and managed the submodule integration, ensuring the update delivered improved performance and reduced the risk of regressions. The upgrade was validated through continuous integration pipelines and a comprehensive pull request workflow, with close collaboration between Kineto and PyTorch maintainers to guarantee seamless adoption. This work enabled more accurate performance measurements and safer deployment of profiling-driven optimizations, reflecting a methodical approach to cross-repository collaboration and robust software integration practices.