AI Architecture 2-Stage Pipeline: Local LLM Generation + Cloud LLM Refinement — Nemotron × Gemini 2.5 Flash Loading...