Tianrui Song

School of Informatics & Edinburgh Futures Institute, The University of Edinburgh

STRAvatar.jpg

Filmed in 2023, Macao, China

Hi! I’m Tianrui Song (feel free to call me Ray), a first-year PhD student in the UKRI CDT in Designing Responsible NLP at the University of Edinburgh, supervised by Prof. Frank Keller. I received my MRes in AI and Machine Learning from Imperial College London (supervised by Prof. Oana-Maria Camburu) and my MPhil in AI from HKUST (Guangzhou) (supervised by Prof. Hao Liu and Prof. Jingshen Wu).

My research interests include Large Language Models, Multimodal LLMs, Agentic AI, and Reasoning, with a focus on AI for creative video production. As a long-time Vlogger, I have a first-hand understanding of the practical challenges creators face, which is what drew me to this direction. My PhD aims to build creator-centric AI assistants that empower, not replace, video creators, and to see them deployed in the real world, genuinely helping myself and fellow creators bring their ideas to life.

Previously, I worked on LLM safety alignment, LLM-enhanced recommender systems, and LLM agents, and interned at ByteDance (Douyin) on recommender systems.

news

Sep 14, 2026 🚚 Moved from London to Edinburgh and officially started my PhD at the University of Edinburgh! 🏰
Jul 22, 2026 🎓 Received a fully funded PhD offer from the University of Edinburgh in the UKRI CDT in Designing Responsible NLP! See everyone in Edinburgh 🏴󠁧󠁢󠁳󠁣󠁴󠁿!
May 15, 2026 💼 Started my internship at ByteDance as an Algorithm Engineer in the Data-Douyin team, working on recommender system iteration!
Mar 15, 2026 Our paper “Mitigating Delivery Fraud and Path Manipulation in UAV-Based E-Commerce: A Fair Exchange Protocol” has been accepted to the IEEE TIFS! 🥳🎉
Nov 18, 2025 🥳 Received the AAAI 2026 Student Scholarship (1000💲 + 500💲)! See everyone in Singapore 🇸🇬!
Nov 08, 2025 Our paper “Hard vs. Noise: Resolving Hard-Noisy Sample Confusion in Recommender Systems via Large Language Models” is accepted by AAAI 2026! 🎊 🥳
Sep 18, 2025 Our paper “MM-Agent: LLM as Agents for Real-world Mathematical Modeling Problem” has been accepted to the NeurIPS 2025! 🥳🎉

selected publications

  1. IEEE TIFS
    IEEE.png
    Mitigating Delivery Fraud and Path Manipulation in UAV-Based E-Commerce: A Fair Exchange Protocol
    Xiaohan Hao, Tianrui Song, Chao Lin, and Xinyi Huang
    IEEE Transactions on Information Forensics and Security, 2026
  2. NeurIPS 2025
    NIPS2025.png
    MM-Agent: LLM as Agents for Real-world Mathematical Modeling Problem
    Fan Liu, Zhe-Rui Yang, Cancheng Liu, Tianrui Song, Xiaofeng Gao, and 1 more author
    In The Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025
  3. AAAI 2026
    AAAI2026.png
    Hard vs. Noise: Resolving Hard-Noisy Sample Confusion in Recommender Systems via Large Language Models
    Tianrui Song, Wen-Shuo Chao, and Hao Liu
    In Proceedings of the AAAI Conference on Artificial Intelligence, 2026