Publications

Where the Blocks Live: Tier-Aware KV-Cache Routing for vLLM at Scale