Worked on the hacksider/Deep-Live-Cam repository to deliver comprehensive GPU-onboarding documentation, focusing on enabling faster GPU-accelerated workloads for new contributors. The main contribution involved updating the README.md using Markdown to include detailed cuDNN installation instructions for CUDA 11.x and guidance on configuring system PATH variables. This documentation clarified the setup process for CUDA and cuDNN, reducing onboarding time and ensuring consistency across development environments. Emphasized documentation best practices and versioned commit traceability, supporting collaboration around GPU-accelerated pipelines. The work established a stronger technical baseline for performance-critical inference and experimentation, directly addressing the needs of GPU-enabled workflows.