Xunliang Cai

Xunliang Cai is a Principal Research Scientist and Director at Meituan, where he leads the AI team and contributes to the LongCat research program. His research spans agentic reinforcement learning, multimodal generation and reasoning, and large language model training and inference. He has co-authored numerous recent publications, including the LongCat-Flash-Omni technical report and several ACL 2026 papers on LLM safety, mid-training, and policy optimization. Before joining Meituan in 2017, he was a Senior Staff Engineer at Baidu.