Worked on the hacksider/Deep-Live-Cam repository to enhance cross-platform machine learning environment consistency by aligning the PyTorch stack on macOS with Linux and Windows. Addressed dependency management by updating requirements.txt, upgrading torch and torchvision to newer releases to ensure compatibility and incorporate security patches. This approach improved reproducibility and reduced environment drift, enabling more reliable ML experiments and streamlining onboarding for new contributors. The work focused on Python and dependency management, leveraging version control to maintain configuration files. The depth of the contribution centered on environment alignment, supporting smoother development workflows across multiple operating systems for machine learning projects.