Design LLM Serving Platform

KV cache, request batching, speculative decoding, and GPU memory scheduling

Estimated time: 15 minutes

Stuck on something? The AI tutor sees this lecture—just ask.