PeopleofSuperintelligence

Researcher

Chulin Xie

Chulin Xie is an AI researcher whose work spans federated-learning security, privacy, trust, and agent safety. She coauthored a distributed backdoor attack in 2020 and VIM, a vertical federated-learning method with user-level privacy. DecodingTrust, which she coauthored in 2023, won a NeurIPS Outstanding Paper Award; she also coauthored an ICML Spotlight paper on private synthetic text in 2024. She became a 2024 Rising Star in Machine Learning, presented on trustworthy foundation models at HKUST in March 2025, and coauthored a study of memorization in language-model reasoning. As of October 2026, she is a research scientist at Google DeepMind in London, working on Gemini, trustworthy agents, and code-agent red teaming.

Timeline

  1. 2020

    A federated backdoor attack

    In 2020, she coauthors DBA, a distributed backdoor attack against federated learning. It divides a global trigger into local patterns that different adversarial parties embed in their training data.

  2. 2022

    Privacy in vertical learning

    In 2022, she and her coauthors introduce VIM, a vertical federated-learning framework that lets clients make multiple local updates before communication. They provide a user-level differential privacy mechanism for the framework.

  3. 2023

    DecodingTrust at NeurIPS

    In 2023, she coauthors DecodingTrust, an evaluation of GPT-4 and GPT-3.5 across toxicity, stereotype bias, adversarial robustness, out-of-distribution robustness, robustness on adversarial demonstrations, privacy, machine ethics, and fairness. NeurIPS gives the paper an Outstanding Paper Award in the Datasets and Benchmarks track.

  4. 2024

    Private synthetic text

    In 2024, she coauthors Differentially Private Synthetic Data via Foundation Model APIs 2: Text, an ICML Spotlight paper. Its Aug-PE method generates differentially private synthetic text through an LLM API without training the model.

  5. 2024

    A 2024 Rising Star

    In 2024, she becomes a Rising Star in Machine Learning. In December 2024, she presents “Assessing and Addressing Trustworthiness Challenges in Language Models” at the University of Maryland’s Rising Stars workshop.

  6. March 2025

    Trustworthy models at HKUST

    In March 2025, she presents “Improving Trustworthiness in Foundation Models: Assessing, Mitigating, and Analyzing ML Risks” at HKUST. Her talk covers safety, privacy, and generalization in foundation models, agents, and federated learning. She is a PhD candidate in computer science at the University of Illinois Urbana-Champaign, advised by Bo Li.

  7. 2025

    Memorization in language reasoning

    In 2025, she coauthors On Memorization of Large Language Models in Logical Reasoning for the IJCNLP–AACL proceedings. The study tests whether high scores on reasoning benchmarks reflect memorization of similar problems.

  8. As of October 2026

    Research at Google DeepMind

    As of October 2026, she is a research scientist at Google DeepMind in London, working on Gemini and trustworthy agents. Her ICML 2026 work, BlueCodeAgent, uses automated red teaming to identify risks in code-generation systems. Her ICLR 2026 work, RedCodeAgent, searches diverse code agents for vulnerabilities.

Career

  1. NowGoogle DeepMindResearch Scientist

Education

  1. University of Illinois Urbana-ChampaignPhD

From the biography.

  1. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  2. Google has the highest combined h-index among AI organizationsIncludes Google. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  3. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Google. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

Sources and credits

Photo: We have no published photo of Chulin Xie yet. Send us a free-licensed photo.

  1. Research | Chulin Xie alphapav.github.io. Checked 22 October 2026.
  2. Improving Privacy-Preserving Vertical Federated Learning by Efficient Communication with ADMM arxiv.org. Checked 22 October 2026.
  3. DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models arxiv.org. Checked 22 October 2026.
  4. Announcing the NeurIPS 2023 Paper Awards blog.neurips.cc. Checked 22 October 2026.
  5. Differentially Private Synthetic Data via Foundation Model APIs 2: Text - Microsoft Research microsoft.com. Checked 22 October 2026.
  6. Rising Stars in Machine Learning Workshop 2024 cml-www.umiacs.io. Checked 22 October 2026.
  7. Improving Trustworthiness in Foundation Models: Assessing, Mitigating, and Analyzing ML Risks cse.hkust.edu.hk. Checked 22 October 2026.
  8. On Memorization of Large Language Models in Logical Reasoning - ACL Anthology aclanthology.org. Checked 22 October 2026.
  9. Chulin Xie alphapav.github.io. Checked 22 October 2026.
  10. BlueCodeAgent: A Blue Teaming Agent Enabled by Automated Red Teaming for CodeGen AI arxiv.org. Checked 22 October 2026.
  11. RedCodeAgent: Automatic Red-teaming Agent against Diverse Code Agents arxiv.org. Checked 22 October 2026.