Over a three-month period, contributed to the ggml-org/llama.cpp repository by developing and refining comprehensive documentation to streamline CUDA toolkit setup on Fedora. Focused on improving onboarding and reducing support overhead, the work included authoring a detailed Fedora CUDA Toolkit Setup Guide, updating it for the 12.8 release, and restructuring content for greater clarity and usability. Leveraging skills in CUDA, Linux, and technical writing with Markdown, the documentation addressed installation complexities, troubleshooting, and performance tuning. All changes were documentation-focused, enhancing the contributor and user experience without modifying code, and demonstrated a methodical approach to supporting GPU-enabled workflows in containerized environments.