Worked on enhancing continuous integration reliability for the pyg-team/pytorch_geometric repository by addressing test stability issues in environments with limited CPU resources. Developed a solution in Python that uses the psutil library to detect the number of physical CPU cores and conditionally skip tests known to fail or behave unpredictably on single-CPU systems. This approach reduced false negatives and flaky test results in CI pipelines, leading to more predictable and efficient release cycles. Focused on system configuration and testing, the work improved feedback speed for developers and ensured that test failures more accurately reflected genuine issues rather than hardware limitations.