Ruixue Ding

Advanced Search Algorithms Specialist, Tongyi Labs

Currently, he is mainly responsible for the RAG algorithm architecture of Tongyi Hundred Refine products and the offline algorithm technology of RAG of other products in Tongyi Labs. 7 years of experience in NLP & AI algorithm research and development as well as landing. He has published many papers in top conferences such as ACL, EMNLP, NAACL, SIGIR, etc. His research area involves NLP traditional tasks, multimodal pre-training, RAG, and he has proposed the industry's first geographic multimodal pre-training model, MGeo, which has been downloaded more than a million times. Currently, he is committed to the construction of process-oriented and modularised landable RAG technology solutions, and has open-sourced the CQDA RAG dataset as well as the CoFE-RAG full-link RAG evaluation framework.