Wei Cui
Principal Researcher
About
Key Areas:
- Cost-efficient inference and low-cost post-training.
- Throughput/latency oriented inference with cloud accelerators and consumer-grade gaming devices/PC.
- Database acceleration on single-GPU and distributed GPU systems.
- Intra-agent and multimodal system optimization.
Other Works:
microsoft/tutel: Tutel MoE: Optimized Mixture-of-Experts Library (opens in new tab)