Hi there, I’m Yuxuan! I am a Machine Learning Engineer working at OpenSearch, AWS. My interests include training sparse and dense embedding models and designing their fine-tuning algorithms. I recently completed my Ph.D. in the Institut des Systèmes Intelligents et de Robotique (ISIR) at Sorbonne University (SU), supervised under Benjamin Piwowarski. Prior to that, I obtained an Engineering Diploma in Computer Engineering from INSA Lyon.

Google Scholar LinkedIn ORCID X (formerly Twitter)

Publications and Blogs

I have authored several papers published in international or french-speaking proceedings, mainly focus on Information Retrieval.

1. MICE to Meet You: From Re-ranking Only to End-to-End Retrieval
   Victor Morand, Yuxuan Zong, Mathias Vast, Josiane Mothe, Benjamin Piwowarski
   Under Review
   pdf code
2. STORM: Stepwise Token Optimization with Reward-Guided Beam Search
   Arthur Satouf, Giulio D’Erasmo, Yuxuan Zong, Habiboulaye Amadou Boubacar, Pablo Piantanida, Benjamin Piwowarski;
   EMNLP 2026: The 2026 Conference on Empirical Methods in Natural Language Processing;
   pdf code
3. From Tokens to Concepts: Leveraging SAE for SPLADE
   Yuxuan Zong, Mathias Vast, Basile Van Cooten, Laure Soulier, Benjamin Piwowarski;
   SIGIR 2026: The 49th International ACM SIGIR Conference on Research and Development in Information Retrieval;
   pdf code
4. A Voronoi Cell Formulation for Principled Token Pruning in Late-Interaction Retrieval Models
   Yash Kankanampati, Yuxuan Zong, Nadi Tomeh, Benjamin Piwowarski, Joseph Le Roux;
   SIGIR 2026: The 49th International ACM SIGIR Conference on Research and Development in Information Retrieval;
   pdf code
5. QueStER: Query Specification for Generative keyword-based Retrieval
   Arthur Satouf, Yuxuan Zong, Habiboulaye Amadou-Boubacar, Pablo Piantanida, Benjamin Piwowarski;
   EACL 2026: The 19th Conference of the European Chapter of the Association for Computational Linguistics;
   pdf code
6. Towards Lossless Token Pruning in Late-Interaction Retrieval Models;
   Yuxuan Zong, Benjamin Piwowarski;
   SIGIR 2025: The 48th International ACM SIGIR Conference on Research and Development in Information Retrieval;
   pdf code
7. Structured representation for Information Retrieval;
   Yuxuan Zong, Benjamin Piwowarski;
   CORIA 2024: COnférence en Recherche d’Informations et Applications;
   pdf
8. Simple Domain Adaptation for Sparse Retrievers;
   Mathias Vast, Yuxuan Zong, Benjamin Piwowarski, Laure Soulier;
   ECIR 2024: The 46th European Conference on Information Retrieval;
   pdf code;
9. XpmIR: A Modular Library for Learning to Rank and Neural IR Experiments;
   Yuxuan Zong, Benjamin Piwowarski;
   SIGIR 2023: The 46th International ACM SIGIR Conference on Research and Development in Information Retrieval;
   pdf code;

Meanwhile, I have also drafted some technical blogs:

1. NVIDIA Nemotron embeddings for agentic search in OpenSearch

Teaching and Academic Service

• Teaching assistant
  • Python, 2022; Sorbonne University; LU1IN001; project design, TP, TP grading
• Program committee member
  • ICTIR 2025
  • SIGIR 2025, 2026
  • ECIR 2027

Invited Talks & Presentation

• Presentations
  • Towards Lossless Token Pruning in Late-Interaction Retrieval Models
    • At MLIA, ISIR: 12/05/2025, post
    • At CORIA-TALN: 02/07/2025, post
    • At SIGIR: 15/07/2026
    • Slides
  • From Tokens to Concepts: Leveraging SAE for SPLADE
    • At CORIA-TALN: 23/06/2026
    • At SIGIR: 23/07/2026
• Invited Talks
  • Towards Efficient and Effective Vocabulary in Sparse Information Retrieval
    • At LIPN: 13/04/2026 (Thanks the invitation from Yash and Tian!)

Resources

Outdated, please check my github for more details

I have participated in creating the following resources:

• XpmIR providing components for launching basic IR experiments, based on experimaestro.
• a code repository providing the experimenting and reproducing scripts for the work Simple Domain Adaptation for Sparse Retrievers, under the framework of XpmIR.

I have also contributed to key open-source toolkits in IR, including:

• ir-datasets, a python toolkit for accessing the dataset of the IR tasks.
• datamaestro_text, a python toolkit for prepare the data record for IR, especially used for XpmIR.

Education

• October 2022 - March 2026
  • Ph.D in Artificial Intelligence, Sorbonne University, France
  • Supervisor: Benjamin Piwowarski
• September 2021 - Febuary 2022
  • Master Exchange in Computer Science, EPFL, Switzerland
• September 2017 - July 2022
  • Engineer Diploma in Computer Engineering, INSA LYON, France

Experience

• October 2026 - Present
  • Machine Learning Engineer, OpenSearch, AWS
• March 2022 - August 2022
  • Research Intern
  • Lab LIRIS, École Centrale de Lyon, France
  • Working on the data and study the state-of-the-art of 6D pose estimation
  • Supervisor: Liming Chen
  • Report
• June 2021 - August 2021
  • Backend software development engineer intern
  • Departement TEG, Tencent, Shenzhen, China
  • Building configuration of the server, based on Python
  • Mentor: Xiangtao LI
• June 2020 - August 2020
  • Frontend software development engineer intern
  • Migu, Nanjing, China
  • Building a mobile application destinate at the companies.

Contact

• Working email:
  • yuxuan.zong AT isir.upmc.fr (deprecated soon sadly >3)
  • beauzong AT amazon.com
• Personal email: zongyuxuaner AT gmail.com