llama.cpp Workflow: Download, Convert, Quantize, and Serve Models