OneRec Technical Report
Formulates recommendation as end-to-end sequence generation and studies how performance scales with training compute.
Researcher at Tencent
杨舒杰
I’m a researcher on Tencent’s WeChat Channels team. My work spans large language models, AI agents, and generative recommendation. I’m interested in how models learn from user behavior and feedback, and how reinforcement learning can improve recommendation and agent systems.
Previously, I worked on OneRec at Kuaishou, focusing on reinforcement learning from human feedback for generative recommendation. I earned my M.S. at Tsinghua University, where I applied machine learning to photonics problems such as metamaterial inverse design and few-shot spectrum inference.
Formulates recommendation as end-to-end sequence generation and studies how performance scales with training compute.
Introduces a decoder-only recommender that uses feedback from deployed systems to improve alignment with user preferences.
Uses a contrastively pretrained language model to design metamaterials that operate across multiple configurations.
Evaluates whether language models can infer fine-grained interests from long histories of user behavior.
* Indicates equal contribution.
Cauchyformer and CoSP were published online in ACS Omega and Advanced Science, respectively.
Joined Tencent’s WeChat Channels team.
The ALPBench preprint on long-term preference modeling appeared on arXiv.
Released the OneRec-V2 Technical Report.
Released the OneRec Technical Report and joined Kuaishou’s Recommendation LLM Group as a research intern.
Joined Xiaohongshu’s AIGC Group as a research intern.
AHEAD received the Best Paper Award at CIAC 2023.
Mar 2026 – Present
Researcher, Tencent Qingyun Program. Large language models, AI agents, and generative recommendation.
Jun 2025 – Mar 2026
Research intern. Worked on reinforcement learning from human feedback for OneRec. Mentored by Qigen Hu.
Nov 2024 – Mar 2025
Research intern. Worked on text-to-image generation with Stable Diffusion and FLUX. Mentored by Shuang Sun.
Aug 2023 – Jun 2026
M.S. in Data Science and Information Technology, Tsinghua-Berkeley Shenzhen Institute. Advised by Kaichen Dong and Yansong Tang.
Aug 2019 – Jul 2023
B.E. in Information and Computational Science, School of Mathematics and Statistics. Advised by Minnan Luo.
Feb – May 2023
Visiting student in the VIP Group. Worked on industrial image anomaly detection with Jinbao Wang.
Outside work, I enjoy running and playing badminton.