Likun Ni

Spatial Data Scientist & AI Engineer
nilk2022@mail.sustech.edu.cn | connect@fuyune.com | +86 132-6557-6569

Professional Summary

Data Science student with a solid foundation in statistics and hands-on experience in geospatial analysis, LLM agent development, NLP, and network science. Proven ability to build end-to-end data pipelines, evaluate spatial reasoning models, and apply complex stochastic and geographic modeling to real-world datasets. Multilingual (Chinese, English, Japanese) with cross-functional research and internship experience in Japan and Hong Kong.

Technical Skills

Python (NumPy, Pandas, PyTorch, NetworkX) R & Java Geospatial Analysis (QGIS) LLM/LMM Applications Network Science & Data Pipelines Software (LaTeX, Gephi) Japanese (JLPT N1) English (IELTS 6.5)

Work Experience

LocationMind Inc. | Tokyo, Japan Jul 2025 – Mar 2026
Data Engineering & Spatial Analysis Intern
  • Conducted extensive literature analysis on geospatial data applications and predictive mobility modeling.
  • Analyzed multi-dimensional geographic datasets, designing robust data pipelines for visualization and reporting.
  • Engineered location-based features for commercial POIs by integrating urban, socioeconomic, and infrastructure data to build comprehensive geographic profiles.
  • Evaluated spatial reasoning capabilities of Large Multimodal Models (LMMs) through prompt engineering, fusing satellite imagery with structured geographic contexts.
Tech Stack: Python, Pandas, GeoPandas, QGIS, LMM APIs (e.g., GPT-4V)
Southern University of Science and Technology Sep 2025 – Dec 2025
Teaching Assistant - Prescriptive Decision Analytics (MIS204)
Instructor: Asst. Prof. Liyi Gu
  • Tutored students on core operations research topics, including linear programming, network models, inventory theory, and decision analysis.
  • Assisted the instructor in grading assignments and quizzes, ensuring clear communication of complex mathematical concepts.

Research & Project Experience

The University of Tokyo | Tokyo, Japan Feb 2025 – Aug 2025
Visiting Research Assistant - Social-Interaction-Aware LLM Agents
Advisor: Prof. Hiroki Hill Kobayashi
  • Designed a tool-calling LLM agent architecture to retrieve historical trajectory data and generate plausible, individualized mobility sequences.
  • Built a geospatial Data Converter tool utilizing point-in-polygon spatial joins to embed Japan level-3 administrative region descriptions into raw user trajectory coordinates.
  • Developed "Preference Statistics" and "Trend Context" modules to extract long-term geographical preferences and short-term behavioral trends from hourly spatiotemporal data.
  • Integrated GIS-based distance calculation tools to verify constraints and improve the spatial plausibility of the generated mobility sequences.
Tech Stack: Python, LangChain, GeoPandas, Shapely, LLM APIs
University of Hong Kong Jul 2024 – Sep 2024
NLP Researcher - Entity Extraction & Collaboration Network Construction
Mentor: Asst. Prof. Shihui Feng
  • Applied Large Language Models (LLMs) to extract entities and relationships from large-scale COVID-19-related news.
  • Constructed organization-level collaboration networks and optimized data processing workflows.
Tech Stack: Python, Hugging Face, NetworkX, Gephi
Southern University of Science and Technology Jan 2024 – Present
Research Project - Concept Embedding and Author Network Analysis
Mentor: Asst. Prof. Yifang Ma
  • Developed concept embedding methods to analyze academic domains and author performance.
  • Explored relationships between author features and their academic impact using vector representations.
Tech Stack: Python, PyTorch, Word2Vec, NetworkX
Southern University of Science and Technology May 2024 – Jun 2024
Course Project - The Small-World Phenomenon under a Restricted Kleinberg Model
Mentor: Assoc. Prof. Yanging Hu
  • Introduced directional constraints in Kleinberg's model to examine connectivity patterns in small-world networks. Co-authored course paper.
Tech Stack: Python, NetworkX, NumPy

Education

National University of Singapore (NUS) Expected Aug 2026
M.Sc. in Data Science for Sustainability
Southern University of Science and Technology (SUSTech) 2022 – 2026
B.S. in Statistics | GPA: 3.49/4.00
Advisor: Asst. Prof. Yifang Ma
Relevant Coursework: Probability Theory, Stochastic Processes, Mathematical Statistics, Linear Algebra, Real Analysis, Network Science, Computing

Publications & Awards

  • Publication: Yukun Jiang, Chongjun Ma, Zhenzhen Xu, Dongliang Mu, Likun Ni, and Hiroki Kobayashi. "MACA-PA: A Modular Agent-based Clinical Assistant for Perioperative Anesthesia." Proceedings of HIDA 2026, ACM. (Accepted)
  • Publication: Yukun Jiang, Chongjun Ma, Likun Ni, et al. "CAKE: Clinicaltrial-oriented Agent for Knowledge Extraction from Health Records." SJTU 2025 Interdisciplinary Graduate Academic Symposium. (Oral & Poster)
  • Award (2024): Third Prize (Provincial Level), Contemporary Undergraduate Mathematical Contest in Modeling.
  • Award (2023): Third Prize (Provincial Level), National College Student Statistical Modeling Competition.

倪李琨

空间数据科学家 / AI 工程师
nilk2022@mail.sustech.edu.cn | connect@fuyune.com | +86 132-6557-6569

个人简介

拥有扎实的统计学基础,并在空间数据分析、LLM Agent 开发、自然语言处理(NLP)及网络科学领域具备丰富的实践经验。熟练构建端到端的数据处理管道,能有效评估空间推理模型,并擅长将复杂的随机建模和地理空间模型应用于真实的工业与商业数据集。具备多语种优势(中文、英语、日语 JLPT N1),在日本及香港拥有跨领域科研与企业实习经历。

专业技能

Python (NumPy, Pandas, PyTorch, NetworkX) R & Java 空间数据分析 (QGIS) 大语言模型 (LLM/LMM 应用) 网络科学与数据工程 软件工具 (LaTeX, Gephi) 日语 (JLPT N1) 英语 (IELTS 6.5)

工作与实习经验

LocationMind Inc. | 日本东京 2025年7月 – 2026年3月
数据工程与空间分析实习生 (研发部)
  • 对地理空间数据应用和预测性移动建模进行了详尽的文献分析与调研。
  • 分析多维地理数据集,设计并实现用于数据可视化和商业报告的自动化数据管道。
  • 融合城市、社会经济和基础设施数据,为商业 POI(兴趣点)提取并丰富基于位置的空间特征,构建多维地理画像。
  • 通过提示词工程(Prompt Engineering)融合卫星图像与结构化地理上下文,有效评估大型多模态模型 (LMM) 在空间推理方面的能力。
技术栈: Python, Pandas, GeoPandas, QGIS, LMM APIs, 本地部署LLM
南方科技大学 2025年9月 – 2025年12月
助教 - 规范性决策分析 (MIS204)
讲师:Asst. Prof. Liyi Gu
  • 辅导学生学习运筹学核心主题,包括线性规划、网络模型、库存理论和决策分析。
  • 协助讲师批改作业和测验,清晰解答复杂的数学与统计建模问题。

项目与科研经验

东京大学 (The University of Tokyo) | 日本东京 2025年2月 – 2025年8月
访问科研助理 - 基于 LLM Agent 的个人移动性生成
导师:Prof. Hiroki Hill Kobayashi
  • 设计基于工具调用 (Tool-calling) 的 LLM Agent 架构,检索历史轨迹并生成合理且高度个性化的人类移动序列。
  • 利用点在多边形内 (point-in-polygon) 的空间连接技术开发数据转换工具,将日本三级行政区划描述精准嵌入原始用户轨迹坐标中。
  • 开发“偏好统计”和“趋势上下文”模块,从每小时的时空数据中提取长期的地理偏好和短期的行为趋势。
  • 集成基于 GIS 的距离计算工具以验证空间约束条件,显著提高 AI 生成移动序列的空间合理性与逻辑性。
技术栈: Python, LangChain, GeoPandas, Shapely, LLM APIs
香港大学 (University of Hong Kong) 2024年7月 – 2024年9月
科研助理 - 基于 LLM 的实体提取与协作网络构建
导师:Asst. Prof. Shihui Feng
  • 应用大语言模型 (LLM) 从大规模新冠疫情相关的新闻数据集中提取复杂的实体与关系。
  • 构建并可视化组织级别的协作网络,优化并落地端到端的数据处理工作流。
技术栈: Python, Hugging Face, NetworkX, Gephi
南方科技大学 2024年1月 – 至今
科研项目 - 概念嵌入与作者网络分析
导师:Asst. Prof. Yifang Ma
  • 开发概念嵌入方法,用于分析学术领域及作者表现。
  • 利用向量表示探索作者特征与其学术影响力之间的关系。
技术栈: Python, PyTorch, Word2Vec, NetworkX
南方科技大学 2024年5月 – 2024年6月
课程项目 - 受限 Kleinberg 模型下的小世界现象
导师:Assoc. Prof. Yanging Hu
  • 在 Kleinberg 模型中引入方向性约束,以研究小世界网络中的连通性模式(完成课程论文)。
技术栈: Python, NetworkX, NumPy

教育背景

新加坡国立大学 (NUS) 预计 2026年8月 入学
可持续发展数据科学 理学硕士 (M.Sc. in Data Science for Sustainability)
南方科技大学 (SUSTech) 2022 – 2026
统计学 理学学士 | GPA: 3.49/4.00
导师:Asst. Prof. Yifang Ma
相关课程: 概率论,随机过程,数理统计,线性代数,实分析,网络科学,计算科学

学术发表与荣誉奖项

  • 学术发表: Yukun Jiang, Chongjun Ma, Zhenzhen Xu, Dongliang Mu, Likun Ni, and Hiroki Kobayashi. "MACA-PA: A Modular Agent-based Clinical Assistant..."(已被 ACM HIDA 2026 接收)。
  • 学术发表: Yukun Jiang, Chongjun Ma, Likun Ni, 等. "CAKE: Clinicaltrial-oriented Agent..."(于上海交大 2025 交叉学科研究生学术论坛进行口头报告与海报展示)。
  • 竞赛奖项 (2024): 全国大学生数学建模竞赛 省级三等奖。
  • 竞赛奖项 (2023): 全国大学生统计建模竞赛 省级三等奖。