My Homepage

May be an AI Researcher.

Greetings!

Zilin Xiao received his CS PhD at Rice University, where he was fortunately advised by Prof. Vicente Ordonez.

His research broadly focuses on vision-language models, large language models, and their reasoning & embedding capabilities -- especially how reasoning and retrieval can be mutually beneficial for each other.

His recent research interests lie in:

  • Vision-Language Models (VLMs): grounding, alignment, and reliable multimodal reasoning
  • Large Language Models (LLMs): post-training for reasoning-centric models
  • Embeddings & Retrieval: retrieval, re-ranking and representation learning

[GitHub] GitHub followers [Personal Email] [Calendly Schedule]


News

[2026.04] Zilin successfully defended his PhD thesis Teaching Vision-Language Models to Retrieve and Reason!

[2026.01] MetaEmbed and ProxyThinker are accepted to ICLR 2026!

[2025.10] We present MetaEmbed, the first work that introduce test-time scaling in multi-vector retrieval.

[2025.05] We present ProxyThinker, a training-free paradigm that steers a VLM to behave like a RL slow-thinking model.

[2025.02] LOCORE is accepted to CVPR 2025! We are the first to use long-context sequence model to do list-wise re-ranking at the local feature level.

[2025.02] Zilin will join Meta GenAI as a research intern in 2025 Summer!

[2024.09] ReHEAT is selected as oral spotlight presentation at COLM 2024!

[2024.07] ReHEAT is accepted to COLM 2024. We explore the connection between language models and a famous psychological cognitive trap called representativeness heuristic.

[2024.07] AutoVER is accepted to ECCV 2024. We augment modern multi-modal language model with precise entity retrieval ability w/ guided generation.

[2024.05] Zilin joined Tencent AI Lab Seattle as a research intern.

[2023.10] Two first-authored papers are accepted to EMNLP 2023 (InsGENEL for Main and CoherentED for Findings).

[2023.08] Zilin is now a first-year PhD student at Rice.

[2022.08] Zilin joined EdgeML Team @ Microsoft as a gap-year research intern.


Education

Rice University
Ph.D. in Computer Science with Prof. Vicente Ordóñez Román
Aug 2023 - May 2026

Sichuan University
B.Eng. in Computer Science with Prof. Yanan Sun
Sep 2018 - Jun 2022


Selected Experience

Meta Superintelligence Labs (MSL)
Research Internship with Dr. Qi Ma at MSL Search, May 2025 - Mar 2026
Focus on scaling multimodal retrieval at test-time with VLMs.

Tencent AI Lab Seattle
Research Internship with Dr. Hongming Zhang and Dr. Tao Ge , May. 2024 - Aug. 2024
Focus on efficient decoding for large language models.

Microsoft
Residency Research Internship with Dr. Ming Gong and Linjun Shou , Aug. 2022 - Jun. 2023
Focus on knowledge-aware and topic-driven entity linking with retrieval-augmented language models.


Awards & Service

National Scholarship Top 0.2% nationwide, 2019

National Scholarship Top 0.2% nationwide, 2020

Outstanding Undergraduate Thesis of Sichuan University 2022.06

Star of Computer Science, Sichuan University 2022.06

Reviewer: EMNLP, ACL ARR, NeurIPS, ICLR, AISTATS, ICML

Teaching Assistant: 2024 Fall COMP 554 - Reinforcement Learning


Hobbies (Click to Expand)

Blockchain: Interested in blockchain technologies and their long-term impact on global finance (Solidity, cryptocurrencies, DeFi).

Trading: Follow markets closely and enjoy turning observations into profit.

Law: Familiar with criminal law in Mainland China.

Flight Simulator: Fly the Airbus A320neo in Microsoft Flight Simulator.

Racing 🏎: Track-day enthusiast. Spend time on tracks in California and Texas (Thunderhill East, Circuit of the Americas, MSR Houston, and growing...).