My research interests lie in accelerating and optimizing systems for AI applications, particularly in large-scale, multi-region, and multi-cloud environments. I am a core contributor to SkyPilot.
May '30  
We arxived MORI, exploiting tool-call idle windows for memory offloading in agentic LLM serving!
May '18  
Excited to join ByteDance Seed as a summer intern — grab a coffee with me if you are in the South Bay!
Jan '10  
We arxived SkyNomad, a multi-region spot scheduler for AI batch jobs to reduce cost by up to 4x!
Idleness is Relative: Exploiting Tool-Call Idle Windows for Offloading in Agentic Systems with MORI (PDF)
Tian Xia, Hanchen Li, Zhifei Li, Xiaokun Chen, Hao Kang, Yifan Qiao, Yi Xu, Ion Stoica.
arXiv 2026.
Agentic Systems; LLM Serving; Memory Offloading; Tool Calls.
SkyNomad: On Using Multi-Region Spot Instances to Minimize AI Batch Job Cost (PDF)
Zhifei Li*, Tian Xia*, Ziming Mao, Zihan Zhou, Ethan J. Jackson, Jamison Kerney, Zhanghao Wu, Pratik Mishra, Yi Xu, Yifan Qiao, Scott Shenker, Ion Stoica.
arXiv 2026.
AI Batch Jobs; Multi-Region; Spot Instances; Cost Optimization.
SkyWalker: A Locality-Aware Cross-Region Load Balancer for LLM Inference (PDF)
Tian Xia, Ziming Mao, Jamison Kerney, Ethan J Jackson, Zhifei Li, Jiarong Xing, Scott Shenker, Ion Stoica.
EuroSys 2026.
Load Balancing; AI Serving; Multi-Region; Cloud Computing.
SkyServe: Serving AI Models across Regions and Clouds with Spot Instances (PDF)
Ziming Mao*, Tian Xia*, Zhanghao Wu, Wei-Lin Chiang, Tyler Griggs, Romil Bhardwaj, Zongheng Yang, Scott Shenker, Ion Stoica.
EuroSys 2025.
Spot Instance; AI Serving; Multi-cloud; Cloud Computing.