Bio

Xiyuan Yang (杨希渊) is now an undergraduate (sophomore) in School of Artificial Intelligence, Shanghai Jiao Tong University (SJTU-SAI). During his freshman year of study, he ranked SECOND of 62 in grades and received honors including the National Scholarship (5 scholarships in total). His research interests focus on Digital Agents with broad capability boundaries (Agentic Scaling Law) and Collaborative Multi-agent Intelligence. He joined the research group MAGIC in SJTU, under the supervision of Prof. Siheng Chen.

Education

School of Artificial Intelligence, Shanghai Jiao Tong University.
GPA: 4.1/4.3 (Ranked 2 out of 62)
Score: 94.0/100
Scholarships: National Scholarship (first 3%), "Han Ying Ju Hua" Scholarship (15 per year), Zhiyuan Honor Scholarships (first 50%), SJTU Undergraduate Excellence Scholarship
High Graded Courses:

• Comprehensive Programming Practice: 100/100
• Probability and Statistics (Honor): 100/100
• Algorithm Design and Analysis: 100/100
• Numerical Analysis: 100/100
• Linear Algebra (Honor): 98/100
• Fundamentals of Programming (Honor): 98/100
• Fundamentals of ML, DL and RL: 98/100

Research

My research interests focus on exploring the Agentic Scaling Law for large language models in the Agent era, to expand the capability boundaries of language model agents in areas such as tool calling, reasoning generalization, and long-term memory. I also investigate multi-agent collaboration workflows and information optimization within collective intelligence. I believe the future world will be a world of agents.

A list of topics that I am interested in at the moment are:

• Agentic Tool Calling and Agentic Memory Enhancement
• Multi-agent Systems and Multi-Agent Collaborations.
• Domain-augmented Agents. (Software Coding Agents, etc.)

Publications

InfoMosaic-Bench: Evaluating Multi-Source Information Seeking in Tool-Augmented Agents

ICLR 2026 Accepted

Yaxin Du, Yuanshuo Zhang, Xiyuan Yang, Yifan Zhou, Cheng Wang, Gongyi Zou, Xianghe Pang, Wenhao Wang, Menglan Chen, Shuo Tang, Zhiyu Li, Feiyu Xiong, Siheng Chen

| | |

We introduce InfoMosaic-Bench, the first benchmark dedicated to multi-source information seeking in tool-augmented agents. Covering 6 representative domains (medicine, finance, maps, video, web, and multi-domain integration), InfoMosaic-Bench requires agents to combine general-purpose search with domain-specific tools. Tasks are synthesized with InfoMosaic-Flow, a scalable pipeline that grounds task conditions in verified tool outputs, enforces cross-source dependencies, and filters out shortcut cases solvable by trivial lookup. This design guarantees both reliability and non-triviality.

AppCopilot: Toward General, Accurate, Long-Horizon, and Efficient Mobile Agent

Jingru Fan, Yufan Dang, Jingyao Wu, Huatao Li, Runde Yang, Xiyuan Yang, Yuheng Wang, Chen Qian

|

We introduce AppCopilot, a multimodal, multi-agent mobile agent designed for seamless cross-app operation. It implements a complete end-to-end pipeline encompassing data collection, model training, fine-tuning, efficient inference, and deployment across PC and mobile platforms. At the model level, it integrates multimodal foundation models with robust bilingual (Chinese-English) support. The reasoning and control layer employs chain-of-thought reasoning, hierarchical task decomposition, and multi-agent collaboration. The execution layer supports experiential adaptation, voice commands, function calling, cross-app and cross-device orchestration, and full mobile app compatibility. This work provides a concrete roadmap and actionable reference architecture for building general-purpose mobile agents.

Project

Skills

Programming Languages

Tools

Python Modules

ML & DL

• Deep learning architectures
• Model training techniques
• Reinforcement learning

LLM & Agents

• Large Language Models
• Post training techniques
• Multi-agent frameworks

Languages

• Chinese (native)
• English (CET-6: 584)

Technical Blogs

|

Active Github committer with 30+ open-source repositories and 1500+ commits.

Maintainer of Xiyuan Yang's Technical Blog. I regularly publish technical content focusing on computer science and AI. To date, I have authored over 120 articles with a cumulative word count exceeding 450,000 words.

Technical Projects

More will be added here in the future. You can see Tool Zoo for more info.

IntelliSearch V3.1: Unifying Search, Empowering Action for Tool Calling Autonomous Agents.

We introduce IntelliSearch, a search-augmented intelligent agent for reliable decision support in real-world, high-noise and multi-constraint environments. Built on the Model Context Protocol (MCP), the system enables large language models to interact with heterogeneous tools, data sources, and execution environments through standardized, verifiable interfaces. IntelliSearch integrates 16 MCP servers and over one hundred carefully engineered tool interfaces, supporting multi-source retrieval, controlled execution, and feedback-driven reasoning. The agent architecture decouples reasoning, memory, and tool scheduling to better support long-horizon tasks and system extensibility. Evaluations on realistic scenarios, including complex transportation planning and cross-source academic retrieval, show that IntelliSearch effectively combines breadth-oriented recall with depth-oriented verification, producing accurate and actionable results in practical settings.

E³-ML-Master: Advanced Envisioning-Executing for Goal-Driven Self-Evolved ML-Master

We propose E³-ML-Master, an agent framework that enhances environment interaction and introspective reasoning through unified Envisioning, Executing, and Evolution. The framework employs a dual-layer loop in which an Envisioner performs global MCTS-based reasoning over long-term memory, while an Executor conducts fine-grained, environment-aware optimizations via tool interactions. Experiments on MLE-Bench show that E³-ML-Master autonomously discovers more efficient strategies and achieves superior performance, while its general architecture provides a scalable foundation for complex, environment-interactive tasks beyond machine learning.

AlphaBuild: Generating Formulaic Alphas on a Wider Range of Stock Data

Leading AlphaBuild as the final project of course SJTU-AI1803, an RL-based methodology using PPO to optimize factor mining process based on AlphaGen, achieving superior backtest performance by leveraging smaller, parallel factor pools and PCA for final factor selection.

SAI Community: The first open-source SAIer's forum for courses, careers and future.

The first open-source learning community and technical exchange platform of the School of Artificial Intelligence at Shanghai Jiao Tong University, featuring course material sharing, internship information dissemination, and domain-specific technical knowledge exchange.

Awards

• 2026-01: "Han Ying Ju Hua" Scholarship (15 per year.)
• 2025-12: SJTU Undergraduate Excellence Scholarship
• 2025-10: Zhiyuan Honor Scholarships (first 50 %)
• 2025-10: Challenge Cup Defender (揭榜挂帅擂台赛 人工智能赛道 擂主 特等奖第一名)
• 2025-09: National Scholarship (first 3%)
• 2025-01: Honorable Mention in MCM 2025
• 2024-10: Zhiyuan Honor Scholarships (first 50 %)

Website Pages Statistics