Portrait
Qingyu Yin
Post-training Lead
Amazon Rufus
About Me

I lead post-training for Rufus, Amazon LLM-powered shopping assistant. As a founding member, I built the in-house foundation models from scratch and build Rufus end-to-end, from core modeling to customer-facing experiences. I currently lead a team of 30 scientists, owning the full post-training stack across SFT and RL. We design and scale post-training recipes to advance core capabilities—reasoning, instruction following, agentic behavior, and helpfulness—while driving application-level preference alignment for Rufus.

Prior to Rufus, I was an Applied Scientist at Query Understanding team (Amazon Search), where I led the development of large-scale query parsing and NLP systems serving millions of customers daily.

News 🔥

  • [Apr. 2026] 3 Papers accepted by ACL 2026! See you in San Diego.

Experience

Amazon — Rufus, Store Foundation Modeling Team
Post-training Lead & Science Manager
Mar. 2023 – Present
  • Founding member of Rufus; trained the in-house LLM from scratch and built Rufus end-to-end, from core model development to customer-facing features, and successfully launched it worldwide.
  • Core Post-Training Recipe: Designed multi-stage training recipe integrating SFT and RL (rubrics, RLVR, agentic training), adopted across multiple Rufus models. Built the first agent optimized for Amazon Search with rule-based RL rewards.
  • Live Traffic Alignment: Developed generative reward models trained on multi-signal live traffic data to align model behavior with real-world customer preferences.
  • Core Feature Development: Partnered with Product team to design and launch core Rufus features worldwide, including product recommendation, comparison, and Product QA.
Amazon — Search, Query Understanding
Applied Scientist
Jul. 2019 – Mar. 2023
  • Senior Applied Scientist owning end-to-end design, development, and deployment of large-scale search and NLP systems serving millions of customers daily.
  • Query Parsing: Led the first large-scale query parsing system at Amazon, enabling explicit attribute extraction to understand customer shopping intent. Adopted by dozens of internal teams with over $1B in incremental revenue impact.

Publications & Preprints

Google Scholar
  • [EMNLP'25] Can Language Models Follow Multiple Turns of Entangled Instructions?
    Chi Han, Xin Liu, Haodong Wang, Shiyang Li, Jingfeng Yang, Haoming Jiang, Zhengyang Wang, Qingyu Yin, et al.
  • [NAACL'25] IHEval: Evaluating Language Models on Following the Instruction Hierarchy
    Zhihan Zhang, Shiyang Li, Zixuan Zhang, Xin Liu, Haoming Jiang, Xianfeng Tang, Yifan Gao, Zheng Li, Haodong Wang, Zhaoxuan Tan, Yichuan Li, Qingyu Yin, Bing Yin, Meng Jiang.
  • [npj Digital Medicine'25] A Multimodal Multidomain Multilingual Medical Foundation Model for Zero Shot Clinical Diagnosis
    Fenglin Liu, Zheng Li, Qingyu Yin, Jinfa Huang, et al.
  • [NeurIPS'24] Shopping MMLU: A Massive Multi-Task Online Shopping Benchmark for Large Language Models
    Yilun Jin, Zheng Li, Chenwei Zhang, Tianyu Cao, Yifan Gao, Pratik Jayarao, Mao Li, Xin Liu, Ritesh Sarkhel, Xianfeng Tang, Haodong Wang, Zhengyang Wang, Wenju Xu, Jingfeng Yang, Qingyu Yin, et al.
    KDD Cup'24.
  • [KDD'24] Understanding Inter-Session Intentions via Complex Logical Reasoning
    Jiaxin Bai, Chen Luo, Zheng Li, Qingyu Yin, Yangqiu Song.
  • [EMNLP'24] Large Language Models in the Clinic: A Comprehensive Benchmark
    Fenglin Liu, Zheng Li, Hongjian Zhou, Qingyu Yin, et al.
  • [NAACL'24] IterAlign: Iterative Constitutional Alignment of Large Language Models
    Xiusi Chen, Hongzhi Wen, Sreyashi Nag, Chen Luo, Qingyu Yin, Ruirui Li, Zheng Li, Wei Wang.
  • [EMNLP'23] Knowledge-Selective Pretraining for Attribute Value Extraction
    Hui Liu, Qingyu Yin, Zhengyang Wang, et al.
  • [EMNLP'23] Improving Consistency for Text Summarization with Energy Functions
    Qi Zeng, Qingyu Yin, Zheng Li, et al.
  • [ACL'23] Context-Aware Query Rewriting for Improving Users' Search Experience on E-commerce Websites
    Simiao Zuo, Qingyu Yin, Haoming Jiang, Shaohui Xi, Bing Yin, Chao Zhang, Tuo Zhao.
  • [ACL'23] Graph Reasoning for Question Answering with Triplet Retrieval
    Shiyang Li, Yifan Gao, Haoming Jiang, Qingyu Yin, Zheng Li, Xifeng Yan, Chao Zhang, Bing Yin.
  • [KDD'23] Knowledge Graph Reasoning over Entities and Numerical Values
    Jiaxin Bai, Chen Luo, Zheng Li, Qingyu Yin, Bing Yin, Yangqiu Song.
  • [NAACL'22] Retrieval-Augmented Multilingual Keyphrase Generation with Retriever-Generator Iterative Training
    Yifan Gao, Qingyu Yin, Zheng Li, Rui Meng, et al.
  • [NAACL'22] All Information Is Valuable: Question Matching over Full Information Transmission Network
    Le Qi, Yu Zhang, Qingyu Yin, Guidong Zheng, Wen Junjie, Jinlong Li, Ting Liu.
  • [NAACL'22] SeqZero: Few-shot Compositional Semantic Parsing with Sequential Prompts and Zero-shot Models
    Jingfeng Yang, Haoming Jiang, Qingyu Yin, Danqing Zhang, Bing Yin, Diyi Yang.
  • [NAACL'22] CERES: Pretraining of Graph-Conditioned Transformer for Semi-Structured Session Data
    Rui Feng, Chen Luo, Qingyu Yin, Bing Yin, Tuo Zhao, Chao Zhang.
  • [EMNLP'21] Logic-level Evidence Retrieval and Graph-based Verification Network for Table-based Fact Verification
    Qi Shi, Yu Zhang, Qingyu Yin, Ting Liu.
  • [ACM TALLIP] Chinese Zero Pronoun Resolution: A Collaborative Filtering-based Approach
    Qingyu Yin, Yu Zhang, Weinan Zhang, Ting Liu.
  • [ACL'19] Towards Explainable NLP: A Generative Explanation Framework for Text Classification
    Hui Liu, Qingyu Yin, William Yang Wang.
  • [ACL'18] Deep Reinforcement Learning for Chinese Zero Pronoun Resolution
    Qingyu Yin, Yu Zhang, Weinan Zhang, Ting Liu, William Yang Wang.
  • [COLING'18] Zero Pronoun Resolution with Attention-based Neural Network
    Qingyu Yin, Yu Zhang, Weinan Zhang, Ting Liu, William Yang Wang.
  • [EMNLP'17] Chinese Zero Pronoun Resolution with Deep Memory Network
    Qingyu Yin, Yu Zhang, Weinan Zhang, Ting Liu.
  • [CCIR'14] Joint Model for Ellipsis Identification and Recovery
    Qingyu Yin, Weinan Zhang, Yu Zhang, Ting Liu.
    Best Student Paper.

Education

  • Harbin Institute of Technology
    Ph.D. in Computer Science — Advisor: Professor Ting Liu
    Sep. 2013 – Jul. 2019
  • Harbin Institute of Technology
    B.S. in Computer Science
    Sep. 2009 – Jun. 2013