Initial: build llama.cpp from source and run LiquidAI LFM2

initial commit