Joy (Jingying) Zeng
I am a Senior Applied Scientist at Microsoft, where my work focuses on post-training for productivity agents, including defining target model behaviors, developing agentic RL and on-policy distillation techniques to post-train models from frontier AI labs, optimizing both response quality and agent inference efficiency, and building data flywheels that power continuous improvement.
My research interests include Large Language Models (LLMs), Natural Language Processing (NLP), and agentic RL. I am always open to research collaborations across academia and industry. Please feel free to reach out!
News
-
2026.10
Our new project, MiniCorp (Office Simulators for Autonomous Enterprise Agents) has launched!
About: MiniCorp is an office simulator where AI agents work together, make decisions,
and respond to market feedback. It provides an environment for studying AI-run companies and a scalable
source of longitudinal and counterfactual enterprise data for agent training.
Explore MiniCorp
Read the paper
Thanks to my collaborator, Zhenwei Dai, advisory committee, and co-authors!
-
2026.09
New preprint: The Tasteful Agent: Measuring and Improving Taste in Long-Horizon Tasks is out!
About: We study the taste of LLM agents: their ability to pick
the better direction at a decision fork before the outcome is visible. We found the best of 14
frontier models scores 59.7% (random = 25%), and the “taste” is trainable via distilling
a hindsight teacher.
Read the paper
Hugging Face
-
2026.08
New paper: Evolving Agents in the Dark (RHO), improving LLM agents from unlabeled past trajectories via self-preference, is accepted at EMNLP 2026.
Experience
2025.11 – Present
Senior Applied Scientist
Microsoft · Mountain View, CA
2024.06 – 2025.10
Applied Scientist — Rufus
Amazon · Palo Alto, CA
2023.05 – 2024.06
Machine Learning Engineer — GenAI
Nextdoor · San Francisco, CA
2022.05 – 2022.08
Machine Learning Intern
Twitter (Cortex) · San Francisco, CA
2021.10 – 2022.04
Statistician / Research Scientist
Stanford University · Palo Alto, CA
Selected Publications
Wenbo Pan, Shujie Liu, Chin-Yew Lin, Jingying Zeng, Xianfeng Tang, Xiangyang Zhou, Yan Lu, Xiaohua Jia (2026).
Evolving Agents in the Dark: Retrospective Harness Optimization via Self-Preference.
accepted by EMNLP 2026.
Pengfei He, Zhenwei Dai, Xianfeng Tang, Yue Xing, Hui Liu, Jingying Zeng, Qiankun Peng, Shrivats Agrawal, Samarth Varshney, Suhang Wang, Jiliang Tang, Qi He (2026).
To trust or not to trust: Attention-based Trust Management for LLM Multi-Agent Systems.
ACL 2026, long paper.
Yingqian Cui, Zhenwei Dai, Pengfei He, Bing He, Hui Liu, Zhan Shi, Xianfeng Tang, Jingying Zeng, Suhang Wang, Yue Xing, Jiliang Tang, Benoit Dumoulin (2026).
A Reward-Guided Dual-Phase Framework for Adaptive Inference-Time Reasoning.
Findings of ACL 2026, long paper.
Minhua Lin, Enyan Dai, Hui Liu, Xianfeng Tang, Yuliang Yan, Zhenwei Dai, Jingying Zeng, Zhiwei Zhang, Fali Wang, Hongcheng Gao, Chen Luo, Xiang Zhang, Qi He, Suhang Wang (2026).
How Far Are LLMs from Professional Poker Players? Revisiting Game-Theoretic Reasoning with Agentic Tool Use.
ICLR 2026.
Zhiwei Zhang, Hui Liu, Xiaomin Li, Zhenwei Dai, Jingying Zeng, Fali Wang, Minhua Lin, Ramraj Chandradevan, Zhen Li, Chen Luo, Xianfeng Tang, Qi He, Suhang Wang (2026).
Bradley–Terry and Multi-Objective Reward Modeling Are Complementary.
ICLR 2026.
Zhining Liu, Ziyi Chen, Hui Liu, Chen Luo, Xianfeng Tang, Suhang Wang, Joy Zeng, Zhenwei Dai, Zhan Shi, Tianxin Wei, Benoit Dumoulin, Hanghang Tong (2026).
Seeing but Not Believing: Probing the Disconnect Between Visual Attention and Answer Correctness in VLMs.
ICLR 2026.
Minhua Lin, Hui Liu, Xianfeng Tang, Jingying Zeng, Zhenwei Dai, Chen Luo, Zheng Li, Xiang Zhang, Qi He, Suhang Wang (2026).
How Far Are LLMs from Real Search: Rethinking the Complementary Roles of Search and Learning.
ACM Transactions on Knowledge Discovery from Data (TKDD) 2026.
Jingying Zeng*, Zhenwei Dai*, Hui Liu, Samarth Varshney, Zhiji Liu, Chen Luo, Zhen Li, Qi He, Xianfeng Tang (2025).
Examples as the Prompt: A Scalable Approach for Efficient LLM Adaptation in E-Commerce.
SIGIR 2025 Industry Track.
Yingqian Cui, Pengfei He, Jingying Zeng, Hui Liu, Xianfeng Tang, Zhenwei Dai, Yan Han, Chen Luo, Jing Huang, Zhen Li, Suhang Wang, Yue Xing, Jiliang Tang, Qi He (2025).
Stepwise Perplexity Guided Refinement for Efficient Chain-of-Thought Reasoning in Large Language Models.
Findings of ACL 2025, long paper.
Jie Ren, Zhenwei Dai, Xianfeng Tang, Hui Liu, Jingying Zeng, Zhen Li, Rahul Goutam, Suhang Wang, Yue Xing, Qi He (2025).
A General Framework to Enhance Fine-tuning-based LLM Unlearning.
Findings of ACL 2025, long paper.
Fali Wang, Hui Liu, Zhenwei Dai, Jingying Zeng, Zhiwei Zhang, Zongyu Wu, Chen Luo, Zhen Li, Xianfeng Tang, Qi He, Suhang Wang (2025).
AgentTTS: Large Language Model Agent for Test-time Compute-optimal Scaling Strategy in Complex Tasks.
NeurIPS 2025.
Jie Ren, Zhenwei Dai, Xianfeng Tang, Yue Xing, Shenglai Zeng, Hui Liu, Jingying Zeng, Qiankun Peng, Samarth Varshney, Suhang Wang, Qi He, Charu C. Aggarwal, Hui Liu (2025).
Keeping an Eye on LLM Unlearning: The Hidden Risk and Remedy.
NeurIPS 2025.
Jinning Li, Ruipeng Han, Jingling Zeng, Dachun Sun, Chenkai Sun, Hanghang Tong, Chengxiang Zhai, Boleslaw K. Szymanski, Tarek Abdelzaher (2025).
Learning to Slice: Self-Supervised Interpretable Hierarchical Representation Learning with Graph Auto-Encoder Tree.
KDD 2025, Research Track.
Jinning Li, Ruipeng Han, Chenkai Sun, Dachun Sun, Ruijie Wang, Jingying Zeng, Yuchen Yan, Hanghang Tong, Tarek Abdelzaher (2024).
Large Language Model Guided Disentangled Belief Representation Learning on Polarized Social Graphs.
International Conference on Computer Communications and Networks (ICCCN) 2024.
Gabriel F. T. Variane, Alex Dahlen, Caroline Y. Noh, Jingying Zeng, Elisabeth S. Yan, Julianna S. Kaneko, Marcella S. Gouveia, Krisa P. Van Meurs, Valerie Y. Chock (2023).
Cerebral Oxygen Saturation in Neonates: a Bedside Comparison between Neonatal and Adult NIRS Sensors.
Pediatric Research 2023.
Preprint
Wenbo Pan, Zhichao Liu, Shujie Liu, Jingying Zeng, Chin-Yew Lin,
Xianfeng Tang, Yan Lu, Qi He, Xiaohua Jia (2026).
The Tasteful Agent: Measuring and Improving Taste in Long-Horizon Tasks.
arXiv:2609.25804.
Mingjia Wang, Taiting Lu, Ziwei Dong, Sisong Bei, Jingying Zeng, Runze Liu,
Kaiyuan Lin, Hongxing Pan, Kai Zhang, Yizheng Hou, Yangshoudu Zheng, Chenchen Guo, Weiyuan Meng,
Shubin Lyu, Zhijun Zheng, Dexu Wang, Xinyu Bai, Shurui Qian, Zhangzixin, Mengyu Pan, Guoliang Shi,
Ling Ma, Yifan Yang, Qi He, Yi-Chao Chen, Yincheng Jin, Sung-Liang Chen, Mahanth Gowda (2026).
OmniCAD: A Large-Scale Benchmark for 3D Spatial Reasoning in Robotics Assemblies.
arXiv:2608.22637.
Taiting Lu, Runze Liu, Ziwei Dong, Sisong Bei, Jingying Zeng, Mingjia Wang, Zhenghao Li, Kaiyuan Lin, Yi-Shan Wu, Yangshoudu Zheng, Hongxing Pan, Kai Zhang, Guoliang Shi, Ling Ma, Yifan Yang, Jiaying Lu, Qi He, Sung-Liang Chen, Yi-Chao Chen, Yincheng Jin, Mahanth Gowda (2026).
OmniMech: All-in-one Multimodal Mechanical Benchmark for 3D Reconstruction.
arXiv:2608.05539.
Jingying Zeng, Hui Liu, Zhenwei Dai, Xianfeng Tang, Chen Luo, Samarth Varshney, Zhen Li, Qi He (2025).
Cite Before You Speak: Enhancing Context-Response Grounding in E-Commerce Conversational LLM-Agents.
arXiv:2503.04830.
Service
Reviewer
EMNLP, NeurIPS, ICLR
Area Chair
KDD 2026–2027
Organizing Role
KDD 2027 Local Chair
Miscellaneous
- I enjoy traveling
and photography 
. - I built a Photo Copilot App for fun.