6.2: Inference Serving

Optimize model inference with quantization, batching, caching, and serving architecture patterns.