Runguo Li

Runguo Li

Master's student in Computer Science @ NUS. Interested in LLMs, Multimodal Models, Agents and FinTech.

Hi, I’m Runguo Li (李润果) 👋

I am an incoming Master’s student in Computer Science at the National University of Singapore (NUS) (2026.8 – 2027.7), graduating from Shanghai University of Finance and Economics (SUFE) with a B.S. in Business Analytics (GPA 3.73/4.0, ranked 8/128).

My research and engineering interests center on Large Language Models (LLMs), Multimodal Representation Learning, AI Agents, and their applications in Finance & Content Safety. I enjoy bridging cutting-edge model research with real-world industrial systems.

I am currently working with the Tencent Youtu AI Lab on adversarial data synthesis and unified content moderation models, with the SUFE FinAI Center on the open-source financial reasoning model Fin-R1, and with the Head Office of ICBC (Private Banking Department) on scientist-discovery agents.

Research Interests

  • Large Language Models — SFT / GRPO / PPO, reasoning, distillation, alignment
  • Multimodal Models — vision–language–video embeddings, retrieval, fusion
  • AI Agents & RAG — tool use, planning, agent benchmarks for finance
  • AI for Finance — financial reasoning, agent safety, knowledge graphs
  • Content Safety — multi-label moderation, low-resource & few-shot detection

Selected Highlights

  • 🏆 Fin-R1 — co-developer of China’s first open-source financial reasoning LLM (7B). FinEval score 75.2 (2nd place overall).
  • 🏆 WeMM-Embedding — multimodal embedding model reaching open-source SOTA on MMEB-V2 (0.7523) and UVRB video retrieval (0.686).
  • 🏆 Medusa Moderation Model — unified content-safety model used in production: 92.3% accuracy, −35% false positives, +60% review efficiency, hundreds of thousands QPS.
  • 📝 FinVault — Benchmarking Financial Agent Safety in Execution-Grounded Environments — first co-author, under review at ACL 2026.

News

  • 2026.05 — Launched this personal homepage at runguoli.com. 🎉
  • 2026.01 — Started working with the Head Office of ICBC on scientist-discovery agents.
  • 2025.10 — Joined Tencent Youtu AI Lab as a research intern (LLM safety & multimodal).
  • 2025.08 — FinVault submitted to ACL 2026.
  • 2026.08 — Will join the National University of Singapore for the Master’s program in Computer Science.

中文简介 (Chinese)

我是李润果，本科毕业于上海财经大学商务分析专业（GPA 3.73/4.0，排名 8/128），将于 2026 年 8 月赴新加坡国立大学攻读计算机科学硕士学位。

我专注于大语言模型、多模态表征学习、AI Agent 以及它们在金融科技与内容安全领域的应用。目前在腾讯优图实验室、上财财经 AI 中心、中国工商银行总行私人银行部进行研究与实习工作。

代表性工作包括：开源金融推理大模型 Fin-R1（FinEval 综合分 75.2，全国第二）、多模态 Embedding 模型 WeMM-Embedding（MMEB-V2 / UVRB 双榜开源第一）、统一内容审核模型 Medusa（线上准确率 92.3%）。学术论文 FinVault 作为共同第一作者投稿至 ACL 2026。

如需联系，欢迎邮件：li19107254665@gmail.com。

Get in Touch