tracked author

Chao Jin

Peking University PhD student advised by Xin Jin. Equal-contribution author of MegaScale-MoE and recurring PKU systems author across UltraEP / ReLibra-style MoE training systems. No high-confidence personal X account found in this pass.

2 archived notes X: not-found Homepage

Related Notes

按论文归档时间排序，展示该作者在本站已经出现的材料。

归档 Jun 28, 2026

MegaScale MoE: Large Scale Communication Efficient Training of Mixture of Experts Models in Production

围绕 attention SP、FFN EP、communication overlap 和 compressed communication 重构生产 MoE 训练路径。

待审阅 2505.11432-megascale-moe-communication-efficient-training MoE Systems Distributed Training Training Memory

归档 Jun 02, 2026

UltraEP: Unleash MoE Training and Inference on Rack Scale Nodes with Near Optimal Load Balancing

基于 post gating exact load 逐 microbatch / layer 规划 expert replica 与 token reroute，统一 MoE 训练和 prefill 负载均衡。

待审阅 2606.04101-ultraep-rack-scale-moe-load-balancing MoE Systems Distributed Training Inference Scheduling