PeopleofSuperintelligence

Researcher

Thomas Steinke

Thomas Steinke's research interests include differential privacy and its connections to learning. He completed an MSc thesis at the University of Canterbury in Christchurch in 2011 and a computer science PhD at Harvard in 2016, advised by Salil Vadhan, then worked at IBM Research Almaden from 2016 until joining Google in 2020. With Clément Canonne and Gautam Kamath, he presented The Discrete Gaussian for Differential Privacy at NeurIPS 2020; the algorithm was deployed in the 2020 US Census, and the 2024 Caspar Bowden Award for Outstanding Research in Privacy Enhancing Technologies recognized that paper. Hyperparameter Tuning with Renyi Differential Privacy, with Nicolas Papernot, won an ICLR 2022 outstanding paper award, Privacy Auditing with One (1) Training Run won a NeurIPS 2023 outstanding paper award, and Stealing Part of a Production Language Model won an ICML 2024 best paper award. As of October 2026 he is a Member of Technical Staff at Anthropic, working on safeguards, after Google Brain and DeepMind from 2020 to 2026.

Now at
Anthropic
Works
184
Citations
5,040
h-index
30

Publication numbers from OpenAlex, as of 17 October 2026.

Timeline

  1. 2011–2016

    Canterbury and a Harvard PhD

    In 2011 he completes an MSc thesis, Constructive Notions of Compactness in Apartness Spaces, at the University of Canterbury in Christchurch, New Zealand, his hometown. In 2016 he completes a PhD in computer science at Harvard University, advised by Salil Vadhan, and the dissertation is Upper and Lower Bounds for Privacy and Adaptivity in Algorithmic Data Analysis.

  2. 2016–2020

    IBM Research Almaden

    He works at IBM Research Almaden from 2016 until he joins Google in 2020.

  3. 2020

    The discrete Gaussian

    The Discrete Gaussian for Differential Privacy, with Clément Canonne and Gautam Kamath, is presented at NeurIPS 2020. The algorithm is deployed in the 2020 US Census, protecting the confidential data of more than 300 million people in the United States.

  4. 2022

    An ICLR outstanding paper

    Hyperparameter Tuning with Renyi Differential Privacy, with Nicolas Papernot, receives an ICLR 2022 outstanding paper award. The paper gives privacy guarantees for hyperparameter search within the framework of Renyi differential privacy.

  5. 2023

    Privacy auditing in one run

    Privacy Auditing with One (1) Training Run, with Milad Nasr and Matthew Jagielski, receives a NeurIPS 2023 outstanding paper award. The paper proposes a scheme for auditing differentially private machine learning systems with a single training run.

  6. 2024

    Caspar Bowden and ICML

    With Clément Canonne and Gautam Kamath, he receives the 2024 Caspar Bowden Award for Outstanding Research in Privacy Enhancing Technologies for The Discrete Gaussian for Differential Privacy. Stealing Part of a Production Language Model, with Nicholas Carlini and colleagues, receives an ICML 2024 best paper award. The attack recovers the embedding projection layer of a transformer from API access to production language models such as ChatGPT and PaLM-2.

  7. April 2025

    Senior scientist at DeepMind

    In April 2025 he is a Senior Research Scientist at Google DeepMind in Mountain View, California. His research centers around differential privacy, with a particular focus on foundational definitions, privacy attacks, and connections to machine learning.

  8. As of October 2026

    Safeguards work at Anthropic

    He is a Member of Technical Staff at Anthropic, working on safeguards, after posts at Google Brain and DeepMind from 2020 to 2026. His research interests include differential privacy and its connections to learning.

Sources and credits

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

  1. Thomas Steinke thomas-steinke.net. Checked 16 October 2026.
  2. Thomas Steinke - The Mathematics Genealogy Project mathgenealogy.org. Checked 16 October 2026.
  3. Challenges and Opportunities in Differentially Private Machine Learning - UC Irvine Donald Bren School of Information & Computer Sciences ics.uci.edu. Checked 16 October 2026.
  4. Gautam Kamath and colleagues win 2024 Caspar Bowden PET Award | Cheriton School of Computer Science | University of Waterloo uwaterloo.ca. Checked 16 October 2026.
  5. Announcing the ICLR 2022 Outstanding Paper Award Recipients – ICLR Blog blog.iclr.cc. Checked 16 October 2026.
  6. ICLR 2022 Awards iclr.cc. Checked 16 October 2026.
  7. Announcing the NeurIPS 2023 Paper Awards – NeurIPS Blog blog.neurips.cc. Checked 16 October 2026.
  8. Differential Privacy - NeurIPS 2023 Outstanding Paper: Privacy auditing in just one run differentialprivacy.org. Checked 16 October 2026.
  9. ICML 2024 Awards icml.cc. Checked 16 October 2026.