Skip to content
Rui Song
Portrait of Rui Song

“Respond to what comes; do not anticipate what has yet to come. Stay undistracted in the present; do not dwell on what has passed.”

— Guofan Zeng

“物来顺应，未来不迎，当时不杂，既过不恋”

—— 曾国藩

Incoming M.Sc. student · University of Toronto

Rui Song

I work on multimodal retrieval, agent evaluation, and the data systems that make both usable in production.

I received my B.Sc. in Statistics and Computer Science, with a minor in Finance, from McGill University in 2023. In September 2026 I am starting an M.Sc. in Applied Computing at the University of Toronto.

Before the master’s I was a Data Engineering Developer II on the Data Platform and AI team at Bell Canada, where I built ingestion and transformation frameworks, self-serve pipeline tooling, and production RAG systems. Earlier I was an AI research assistant at McGill, working on recommender fairness and multi-agent reinforcement learning.

Recent papers include MINER on multimodal retrieval, QUACK (EMNLP 2026 Main) on grounding in social-deduction agents, CARE (EMNLP 2026 Findings) on privacy-compliant clinical reasoning, and a survey of discrete diffusion models.

Resume Community profile

News

  1. We have two papers accepted to EMNLP 2026, “QUACK: Questioning, Understanding, and Auditing Communicated Knowledge in Multimodal Social Deduction Agents” and “CARE: Privacy-Compliant Agentic Reasoning with Evidence Discordance”.

  2. Our survey Discrete Diffusion Models: A Unified Framework from Tokenization to Generation is on arXiv.

  3. I concluded three years as a Data Engineering Developer II at Bell Canada.

  4. “MINER: Mining Multimodal Internal Representation for Efficient Retrieval” is on arXiv. I am a co-first author.

  5. I started a new research project focusing on visual document retrieval.

  6. I moved to the Data Platform and AI team at Bell Canada to build a new data platform supporting Data Mesh.

  7. I began working as tech lead to help the team migrate to GCP.

  8. I took 3rd place in Bell’s first internal AI hackathon with a RAG system to help the business team.

  9. Unit Cost officially went into production, supporting the business team in analyzing and optimizing cost management.

  10. I started the Unit Cost project, analyzing unit fibre cost for all fibres at Bell Canada.

  11. I started my job at Bell Canada as a Data Engineer.

  12. I received my B.Sc. from McGill University with Distinction (GPA 3.93).

  13. I joined Professor Xue (Steve) Liu’s CPS Lab as a research assistant and attended the group meetings regularly.

  14. I started a research project on fairness in recommender systems with Dr. Haolun Wu and Sirui Song. They were students of Professor Xue (Steve) Liu.

  15. I started a research project on multi-agent reinforcement learning with Dr. Jikun (Jaxon) Kang. He was a student of Professor Xue (Steve) Liu.

  16. I started the course COMP 597 (Applications of Machine Learning in Real World Systems) at McGill University, a research-oriented course taught by Professor Xue (Steve) Liu. I was impressed by the research projects of the students in the course.

  17. I started my internship as a business analyst at Pan-Pacific Group in Xiamen, Fujian, China.

  18. I started my undergraduate study at McGill University.