vllm 5
- The Four LLM Inference Patterns: Disaggregating Prefill and Decode Economics
- Signal-Decision Architecture: Multi-Dimensional Semantic Routing for vLLM Clusters
- Architecting Production ML Inference: From Microservices to vLLM Clusters
- Graph-Structured Speculative Decoding: Accelerating LLM Inference via DAG Verification
- Mitigating LLM Inference Bottlenecks: Prefill Compute vs Memory-Bound Decode Mechanics