Hi there, this is Shufan Jiang. I actively develop and contribute to open-source projects.

My current research interests include:

• Large Language Models (LLMs)
• Multi-Agent Systems
• Scaling Environments for Agent
• Agent Benchmarking & Evaluation Methodologies

I was previously a research intern at miHoYo LumiNLP, where I worked on Intelligent Game NPCs.

Research

GBQA: A Game Benchmark for Evaluating LLMs as Quality Assurance Engineers

• Authored a paper as the first author (Accepted at ICLR 2026 Workshop);
• Proposed GBQA, an interactive benchmark containing 30 games and 124 verified bugs;
• GBQA emphasizes autonomous bug discovery and quality assurance within the development cycle.

FURINA: A Fully Customizable Role-Playing Benchmark via Scalable Multi-Agent Collaboration Pipeline

Supervised by Prof.Chengwei Qin, The Hong Kong University of Science and Technology, China

• Authored a paper as the co-first author https://arxiv.org/pdf/2510.06800;
• Proposed FURINA-Builder, the first multi-agent collaboration pipeline for automatically constructing fully customizable RP benchmarks at arbitrary scales, targeting real-world challenges;
• Introduced FURINA-Bench, a comprehensive RP benchmark built with FURINA-Builder, which incorporates both established and synthesized test characters in group-chat scenarios, accompanied by fine-grained evaluation criteria. A preliminary analysis demonstrates that it facilitates clearer model separability and supports more robust evaluation.

HAMLET: A Hierarchical and Adaptive Multi-agent Framework for Live Embodied Theatrics

Supervised by Prof.Xuelong Li, Institute of Artificial Intelligence (TeleAI), China Telecom, China

• Authored a paper as the first author (Accepted at ICLR 2026);
• Proposed an multi-agent system HAMLET, capable of generating entire online drama performance autonomously with single-sentence user input;
• Conducted extensive experiments including model training, win-rate evaluation and ablation studies.

LINKs: LLM Integrated Management for 6G Empowered Digital Twin NetworKs

Supervised by Dr.Yue Wu, School of Information Science and Engineering, East China University of Science and Technology, China

• Authored a paper as the first author (Accepted at IEEE VTC 2024);
• Proposed an autonomous network management framework LINKs for urban scenarios, incorporating multi-agent collaboration, a self-reflection workflow, and a full toolset for agent tool calling;
• Conducted ablation studies on multiple experimental results, fine-tuned TimesFM, Google’s pretrained time series prediction model, with LoRA using the HuggingFace PEFT Library, achieving state-of-the-art performance in network traffic prediction.

Project

Island Life From Zero

Supervised by Dr.Ke Fang, Tsinghua Shenzhen International Graduate School, Tsinghua University, China

• Designed an island‑themed social‑simulation game in Unity; published on itch.io.
• Implemented LLM‑driven NPCs with planning, reasoning, memory, and tool‑use.
• Enabled NPC skill evolution through interaction and collaboration.

PoetryChat: LLM‑based Interactive Chinese Poetry Learning Assistant

• Designed structured prompts for text-to-image generation and text interaction tailored to different age groups, and integrated them with LangChain;
• Implemented multimodal RAG module by extracting images from documents and recaptioning them via BLIP2-OPT-6.7B for semantic alignment, and built vector storage for multimodal retrieval;
• Developed web interface based on React, featuring history tracking, theme switching, online searching, day-night UI toggling and RAG file uploading.

Multi‑Granularity Lesion Cell Detection Using Deep Neural Networks

Supervised by Prof. Pietro Liò, Department of Computer Science and Technology, University of Cambridge, UK

• Collected and annotated a dataset of over 2,000 medical cell images across 7 categories from various sources, including Kaggle, HuggingFace, and other open-source platforms;
• Pre-processed datasets using OpenCV-Python, including noise reduction, deduplication, and data augmentation with the Mosaic method to generate diverse training samples;
• Trained object detection models using both CNN-based and Transformer-based network architectures, including YOLO series, Vision Transformer, and DEtection TRansformer. Demo displayed on GitHub.

Publications

* denotes equal contribution

2026

• [ICML 2026] CAMEL-AI. CAMEL: A Framework for Finding the Scaling Laws of Agents.
• [ICLR 2026 Workshop] Shufan Jiang, Chios Chen, Zhiyang Chen. GBQA: A Game Benchmark for Evaluating LLMs as Quality Assurance Engineers.
• [ICLR 2026] Shufan Jiang, Sizhou Chen, Chi Zhang, Xiao-Lei Zhang, Xuelong Li. HAMLET: A Hierarchical and Adaptive Multi-agent Framework for Live Embodied Theatrics.

2025

• [arXiv] Haotian Wu*, Shufan Jiang*, Mingyu Chen, Yiyang Feng, Hehai Lin, Heqing Zou, Yao Shu, Chengwei Qin. FURINA: A Fully Customizable Role-Playing Benchmark via Scalable Multi-Agent Collaboration Pipeline.

2024

• [VTC 2024] Shufan Jiang, Bangyan Lin, Yue Wu, Yuan Gao. LINKs: LLM Integrated Management for 6G Empowered Digital Twin Networks.

Open Source Activities

Datawhale

• Core Maintainer of self‑llm.
• Co‑Leader of hello-agents, covers the design and implementation of agent systems, ranging from fundamental theories to practical applications.
• Co‑Leader of unlock‑deepseek, dedicated to interpreting, expanding, and reproducing DeepSeek’s innovative achievements.

CAMEL-AI

• Integrated BFCL (Berkeley Function Calling Leaderboard) features into camel.
• Co-lead multi-agent tutorials based on camel framework in handy‑multi‑agent.

Others

• Contributed to transformers by fixing triton check to support MXFP4 model inference.
• Contributed to LLaMA‑Factory by resolving compatibility issues with bitsandbytes quantization.

Experience

2025 miHoYo

LLM Algorithm Research Intern, LumiNLP, miHoYo Network Technology Co., Ltd.

2024 Roche

AI-Based Drug Design and App Development Intern, CICoR, Roche R&D Center (China) Ltd.

2023 Bilibili

Development Intern (LLM Direction), Corporate Efficiency Department, Shanghai Bilibili Technology Co., Ltd.

2023 Meituan

Software Engineer Intern, Intelligent and Communication Technology Center, Hanhai Information Technology Co., Ltd.