During January 2025, Haus Xda improved AMD GPU architecture detection in both the ggml-org/llama.cpp and Mintplex-Labs/whisper.cpp repositories. Working in C and C++ with a focus on CUDA and low-level hardware programming, Haus refactored the parsing logic to use gcnArchName for more accurate device identification, replacing less specific values and enhancing logging for maintainability. These changes reduced the risk of mis-identification in CUDA environments and aligned parsing logic across repositories, supporting easier troubleshooting and future updates. The work demonstrated a strong understanding of GPU computing and hardware architecture, delivering targeted, maintainable solutions to real-world device detection challenges.