Ruidong Zhu

Latest

• ExpertPlex: A High-Goodput Disaggregated Serving System for MoE LLMs with Adaptive Persistent Kernels
• ReLibra: Routing-Replay-Guided Load Balancing for MoE Training in Reinforcement Learning
• Epiphron: Resource-Efficient Distributed Key-Value Storage
• dLoRA: Dynamically Orchestrating Requests and Adapters for LoRA LLM Serving