About

I am Jue Zhang (张珏), a Principal Researcher at Microsoft, where I lead research on reliable and self-improving AI agents, AI efficiency, and privacy-preserving AI.

My recent research focuses on enabling AI agents to diagnose failures, learn from experience, and continuously improve their own capabilities. From failures to flywheels, I explore how reliable and efficient agentic systems can evolve beyond static, human-designed workflows.

My research has contributed to Microsoft products including Microsoft 365, Bing, and Azure, resulting in hundreds of millions of dollars in cost savings. I have authored nearly 50 peer-reviewed publications across leading AI/ML conferences and physics journals.

Beyond my individual research, I lead collaborative research initiatives and have mentored 10+ research interns.

Featured Research

Selected Publications

See Google Scholar for my full publication list.

Reliable and self-improving agents

1. ActiveSaddler: Automated Curriculum Learning for Agent Harness Optimization
   Sungho Park, Wonjoong Kim, Jue Zhang, Wook-Shin Han, Pengfei Gao, Chanyoung Park, Yongqiang Yao, Rao Fu, Elsie Nallipogu, Qingwei Lin, Victor Rühle
   arXiv:2610.00906 · arXiv · GitHub
2. AutoSaddler: Automatic Harness Optimization with Durable Updates from Agent Execution Traces
   Sungho Park, Wonjoong Kim, Rongyuan Tan, Jue Zhang, Wook-Shin Han, Pengfei Gao, Chanyoung Park, Yongqiang Yao, Rao Fu, Elsie Nallipogu, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang
   NeurIPS, 2026 · arXiv · GitHub · Publisher
3. Contrastive Attribution in the Wild: An Interpretability Analysis of LLM Failures on Realistic Benchmarks
   Rongyuan Tan, Jue Zhang, Zhuozhao Li, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang
   EMNLP, 2026 · arXiv · GitHub · Publisher
4. A Tale of Two Graphs: Separating Knowledge Exploration from Outline Structure for Open-Ended Deep Research
   Zhuofan Shi, Ming Ma, Zekun Yao, Fangkai Yang, Jue Zhang, Dongge Han, Victor Rühle, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang
   ICML, 2026 · arXiv · Publisher
5. DoVer: Intervention-Driven Auto Debugging for LLM Multi-Agent Systems
   Ming Ma, Jue Zhang, Fangkai Yang, Yu Kang, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang
   ICLR, 2026 · arXiv · GitHub · Publisher
6. From Reasoning to Answer: Empirical, Attention-Based and Mechanistic Insights into Distilled DeepSeek R1 Models
   Jue Zhang, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang
   EMNLP, 2025 · arXiv · GitHub · Publisher

AI efficiency

1. AXIS: Efficient Human-Agent-Computer Interaction with API-First LLM-Based Agents
   Junting Lu, Zhiyang Zhang, Fangkai Yang, Jue Zhang, Lu Wang, Chao Du, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Zhang
   ACL, 2025 · arXiv · Publisher
2. TACO-RL: Task Aware Prompt Compression Optimization with Reinforcement Learning
   Shivam Shandilya, Menglin Xia, Supriyo Ghosh, Huiqiang Jiang, Jue Zhang, Qianhui Wu, Victor Rühle, Saravan Rajmohan
   Findings of ACL, 2025 · arXiv · Publisher
3. Performance Aware LLM Load Balancer for Mixed Workloads
   Kunal Jain, Anjaly Parayil, Ankur Mallick, Esha Choukse, Xiaoting Qin, Jue Zhang, Íñigo Goiri, Rujia Wang, Chetan Bansal, Victor Rühle, Anoop Kulkarni, Steve Kofsky, Saravan Rajmohan
   ACM EuroMLSys Workshop, 2025 · Publisher
4. AutoRAG-HP: Automatic Online Hyper-Parameter Tuning for Retrieval-Augmented Generation
   Jia Fu, Xiaoting Qin, Fangkai Yang, Lu Wang, Jue Zhang, Qingwei Lin, Yubo Chen, Dongmei Zhang, Saravan Rajmohan, Qi Zhang
   Findings of EMNLP, 2024 · arXiv · Publisher
5. LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression
   Zhuoshi Pan, Qianhui Wu, Huiqiang Jiang, Menglin Xia, Xufang Luo, Jue Zhang, Qingwei Lin, Victor Rühle, Yuqing Yang, Chin-Yew Lin, H. Vicky Zhao, Lili Qiu, Dongmei Zhang
   Findings of ACL, 2024 · arXiv · GitHub · Publisher

Privacy-preserving AI and Personalization

1. CI-Work: Benchmarking Contextual Integrity in Enterprise LLM Agents
   Wenjie Fu, Xiaoting Qin, Jue Zhang, Qingwei Lin, Lukas Wutschitz, Robert Sim, Saravan Rajmohan, Dongmei Zhang
   ACL Industry Track, 2026 · arXiv · Publisher
2. EgoMemory: Memory-Augmented Personalized Retrieval for Long-Context Egocentric Video
   Yuanmin Tang, Jue Zhang, Xiaoting Qin, Jing Yu, Meikang Qiu, Gaopeng Gou, Gang Xiong, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Wu
   Findings of ACL, 2026 · Publisher
3. Privacy in Action: Towards Realistic Privacy Mitigation and Evaluation for LLM-Powered Agents
   Shouju Wang, Fenglin Yu, Xirui Liu, Xiaoting Qin, Jue Zhang, Qingwei Lin, Dongmei Zhang, Saravan Rajmohan
   Findings of EMNLP, 2025 · arXiv · GitHub · Publisher
4. MEETING DELEGATE: Benchmarking LLMs on Attending Meetings on Our Behalf
   Lingxiang Hu, Shurun Yuan, Xiaoting Qin, Jue Zhang, Qingwei Lin, Dongmei Zhang, Saravan Rajmohan, Qi Zhang
   HCI+NLP Workshop, 2025 · arXiv · Publisher
5. Reason-before-Retrieve: One-Stage Reflective Chain-of-Thoughts for Training-Free Zero-Shot Composed Image Retrieval
   Yuanmin Tang, Jue Zhang, Xiaoting Qin, Jing Yu, Gaopeng Gou, Gang Xiong, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Wu
   CVPR, 2025 · arXiv · GitHub · Publisher
6. Navigating the Unknown: A Chat-Based Collaborative Interface for Personalized Exploratory Tasks
   Yingzhe Peng, Xiaoting Qin, Zhiyang Zhang, Jue Zhang, Qingwei Lin, Xu Yang, Dongmei Zhang, Saravan Rajmohan, Qi Zhang
   ACM IUI, 2025 · arXiv · Publisher
7. Sharingan: Extract User Action Sequence from Desktop Recordings
   Yanting Chen, Yi Ren, Xiaoting Qin, Jue Zhang, Kehong Yuan, Lu Han, Qingwei Lin, Dongmei Zhang, Saravan Rajmohan, Qi Zhang
   arXiv:2411.08768 · arXiv

Cloud Management and Resource Optimization

1. RCInvestigator: Towards Better Investigation of Anomaly Root Causes in Cloud Computing Systems
   Shuhan Liu, Yunfan Zhou, Jue Zhang, Shandan Zhou, Weiwei Cui, Qingwei Lin, Thomas Moscibroda, Haidong Zhang, Di Weng, Yingcai Wu
   IEEE Transactions on Visualization and Computer Graphics (PacificVIS Journal Track), 2026 · arXiv · Publisher
2. Enabling Autonomic Microservice Management through Self-Learning Agents
   Fenglin Yu, Fangkai Yang, Xiaoting Qin, Zhiyang Zhang, Jue Zhang, Qingwei Lin, Hongyu Zhang, Yingnong Dang, Saravan Rajmohan, Dongmei Zhang, Qi Zhang
   arXiv:2501.19056 · arXiv
3. Towards Workload-aware Cloud Efficiency: A Large-scale Empirical Study of Cloud Workload Characteristics
   Anjaly Parayil, Jue Zhang, Xiaoting Qin, Íñigo Goiri, Lexiang Huang, Timothy Zhu, Chetan Bansal
   ACM/SPEC ICPE, 2025 · arXiv · Publisher
4. Workload Intelligence: Workload-Aware IaaS Abstraction for Cloud Efficiency
   Lexiang Huang, Anjaly Parayil, Jue Zhang, Xiaoting Qin, Chetan Bansal, Jovan Stojkovic, Pantea Zardoshti, Pulkit Misra, Eli Cortez, Raphael Ghelman, Íñigo Goiri, Saravan Rajmohan, Jim Kleewein, Rodrigo Fonseca, Timothy Zhu, Ricardo Bianchini
   ACM/IEEE SC, 2025 · arXiv · Publisher
5. The Vision of Autonomic Computing: Can LLMs Make It a Reality?
   Zhiyang Zhang, Fangkai Yang, Xiaoting Qin, Jue Zhang, Qingwei Lin, Gong Cheng, Dongmei Zhang, Saravan Rajmohan, Qi Zhang
   arXiv:2407.14402 · arXiv · GitHub
6. Snape: Reliable and Low-Cost Computing with Mixture of Spot and On-Demand VMs
   Fangkai Yang, Lu Wang, Zhenyu Xu, Jue Zhang, Liqun Li, Bo Qiao, Camille Couturier, Chetan Bansal, Soumya Ram, Si Qin, Zhen Ma, Íñigo Goiri, Eli Cortez, Terry Yang, Victor Rühle, Saravan Rajmohan, Qingwei Lin, Dongmei Zhang
   ACM ASPLOS, 2023 · Publisher
7. How Different are the Cloud Workloads? Characterizing Large-Scale Private and Public Cloud Workloads
   Xiaoting Qin, Minghua Ma, Yuheng Zhao, Jue Zhang, Chao Du, Yudong Liu, Anjaly Parayil, Chetan Bansal, Saravan Rajmohan, Íñigo Goiri, Eli Cortez, Si Qin, Qingwei Lin, Dongmei Zhang
   IEEE/IFIP DSN, 2023 · Publisher
8. Contextual Self-attentive Temporal Point Process for Physical Decommissioning Prediction of Cloud Assets
   Fangkai Yang, Jue Zhang, Lu Wang, Bo Qiao, Di Weng, Xiaoting Qin, Gregory Weber, Durgesh Nandini Das, Srinivasan Rakhunathan, Ranganathan Srikanth, Qingwei Lin, Dongmei Zhang
   ACM KDD, 2023 · Publisher
9. Spot Virtual Machine Eviction Prediction in Microsoft Cloud
   Fangkai Yang, Bowen Pang, Jue Zhang, Bo Qiao, Lu Wang, Camille Couturier, Chetan Bansal, Soumya Ram, Si Qin, Zhen Ma, Íñigo Goiri, Eli Cortez, Senthil Baladhandayutham, Victor Rühle, Saravan Rajmohan, Qingwei Lin, Dongmei Zhang
   ACM WWW Companion, 2022 · Publisher

AI for Physics

1. Does Lorentz-symmetric design boost network performance in jet physics?
   Congqiao Li, Huilin Qu, Sitian Qian, Qi Meng, Shiqi Gong, Jue Zhang, Tie-Yan Liu, Qiang Li
   Physical Review D, 2024 · arXiv · Publisher
2. An efficient Lorentz equivariant graph neural network for jet tagging
   Shiqi Gong, Qi Meng, Jue Zhang, Huilin Qu, Congqiao Li, Sitian Qian, Weitao Du, Zhi-Ming Ma, Tie-Yan Liu
   Journal of High Energy Physics, 2022 · arXiv · GitHub · Publisher

Neutrino Physics

1. Gravitational clustering of cosmic relic neutrinos in the Milky Way
   Jue Zhang, Xin Zhang
   Nature Communications, 2018 · arXiv · Publisher · Media: Guangming Daily (Chinese)
2. Viability of exact tri-bimaximal, golden-ratio and bimaximal mixing patterns and renormalization-group running effects
   Jue Zhang, Shun Zhou
   Journal of High Energy Physics, 2016 · arXiv · Publisher
3. Distinguishing Dirac/Majorana Sterile Neutrinos at the LHC
   Claudio O. Dib, Choong Sun Kim, Kechen Wang, Jue Zhang
   Physical Review D, 2016 · arXiv · Publisher
4. Improved Statistical Determination of Absolute Neutrino Masses via Radiative Emission of Neutrino Pairs from Atoms
   Jue Zhang, Shun Zhou
   Physical Review D, 2016 · arXiv · Publisher
5. Relic Right-handed Dirac Neutrinos and Implications for Detection of Cosmic Neutrino Background
   Jue Zhang, Shun Zhou
   Nuclear Physics B, 2016 · arXiv · Publisher
6. Confronting Four Zero Neutrino Yukawa Textures with N₂-dominated Leptogenesis
   Jue Zhang
   Physical Review D, 2015 · arXiv · Publisher
7. θ₁₃ and The Flavor Ring
   Jennifer Kile, Michael Jay Pérez, Pierre Ramond, Jue Zhang
   Physical Review D, 2014 · arXiv · Publisher

Honors and awards

• Machine Learning, AI and Data Science (MLADS) Distinguished Contribution Award, Microsoft (2023, 2024, 2026).
• Best Poster Award, 28th International Symposium on Lepton Photon Interactions at High Energies (2017).
• China Postdoctoral Fellowship (First Grade), Peking University (2017).
• CLAS Dissertation Fellowship, University of Florida (2014).

Professional service

• Review Process Co-Chair, ICSE 2027.
• Silver Reviewer Award, ICML (2026).
• Area Chair, ACL Rolling Review.
• Reviewer, NeurIPS, ICML, ICLR, AAAI, and ACL Rolling Review.
• Journal Reviewer, Physical Review D and Journal of High Energy Physics.

Bio

Before joining Microsoft, I worked at JD.com on applied research and product development in wearable-based health monitoring and personalized recommendation systems. I contributed to ABC Health, a WeChat mini program that helped people monitor their health during the COVID-19 pandemic. According to 36Kr (in Chinese), the program reached an estimated 10 million users during its pandemic-era surge.

Before transitioning to AI/ML, I was a physicist with a Ph.D. in High Energy Physics from the University of Florida and a B.S. in Theoretical Physics from the University of Science and Technology of China.