LLM Serving

2024-11-27 Optimizing and Characterizing High-Throughput Low-Latency LLM Inference in MLCEngine
2025-12-04 SuffixDecoding: Extreme Speculative Decoding for Emerging AI Applications