Optimizes AI models for edge deployment through quantization, lazy loading, and memory management. Use when deploying models to resource-constrained environments, mobile devices, or edge computing scenarios. Do not use for cloud deployment, model training, or data preprocessing.