# Raja Gond [[Microsoft Research]] India の研究者。[[@2025__arXiv__TokenWeave - Efficient Compute-Communication Overlap for Distributed LLM Inference|TokenWeave]](MLSys 2026)の筆頭著者かつ Corresponding author([email protected])。テンソル並列推論における計算-通信オーバーラップと融合 AllReduce–RMSNorm カーネルの設計を主導した。(Source: [[@2025__arXiv__TokenWeave - Efficient Compute-Communication Overlap for Distributed LLM Inference]]) ## 関連 - ソース: [[@2025__arXiv__TokenWeave - Efficient Compute-Communication Overlap for Distributed LLM Inference]] - 共著者: [[Nipun Kwatra]] / [[Ramachandran Ramjee]] - 所属: [[Microsoft Research]]