inference-optimization

Optimizing AI inference - quantization, speculative decoding, KV cache, batching, caching strategies. Use when reducing latency, lowering costs, or scaling AI serving.