Hao Kang
4 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
Privatar introduces a scalable, privacy-preserving framework to offload computationally intensive multi-user avatar reconstruction from VR headsets to untrusted local devices, significantly improving user capacity while maintaining strong privacy guarantees.
The paper introduces PithTrain, a compact, agent-native Mixture-of-Experts (MoE) training framework that significantly improves agent-task efficiency compared to existing production stacks.
The paper introduces PithTrain, a compact, agent-native Mixture-of-Experts (MoE) training framework that significantly improves agent-task efficiency compared to existing production stacks.
MORI is a system that optimizes memory allocation for agentic workloads in LLM serving systems by dynamically shifting the partition boundary between GPU HBM and CPU DRAM based on program idleness.
Papers
Idleness is Relative: Exploiting Tool-Call Idle Windows for Offloading in Agentic Systems with MORI
Tian Xia, Hanchen Li, Zhifei Li, Xiaokun Chen +4 more
MORI is a system that optimizes memory allocation for agentic workloads in LLM serving systems by dynamically shifting the partition boundary between GPU HBM and CPU DRAM based on program idleness.