PeopleofSuperintelligence

Researcher

Krishnamurthy Dvijotham

Also known as Dj Dvijotham

Krishnamurthy Dvijotham is a research scientist at Google DeepMind whose work aims at secure and trustworthy AI systems, from verifying neural networks to testing AI agents. After a postdoctoral fellowship at Caltech's Center for Mathematics of Information, he worked at Pacific Northwest National Laboratory in 2016 and 2017, then joined Google DeepMind in 2017 and co-founded its robust and verified AI team. In 2018, he coauthored a paper on scalable verification of deep networks, and in 2024 two papers with him as a coauthor won best paper awards: the ICML paper on stealing part of a production language model and the CVPR paper on rich human feedback for text-to-image generation. He led AI safety and security research at ServiceNow Research from 2024 to 2025. He is a senior staff research scientist at Google DeepMind, where he leads work on secure and trustworthy AI agents.

Badges

Timeline

  1. 2016–2017

    Caltech postdoc, then PNNL

    After a postdoctoral fellowship at the Center for Mathematics of Information at Caltech, he works as a researcher at Pacific Northwest National Laboratory from 2016 to 2017.

  2. 2017–2024

    Researcher at Google DeepMind

    He works as a researcher at Google DeepMind from 2017 to 2024. He co-founds the robust and verified AI team at DeepMind, and his research is deployed in several Google products, including the Android Play Store, YouTube and Gemini.

  3. 2018

    Scalable verification of deep networks

    In 2018, he coauthors a paper with Robert Stanforth, Sven Gowal, Timothy Mann and Pushmeet Kohli on verifying properties of neural networks. The method formulates verification as an optimization problem and solves a Lagrangian relaxation to bound the worst-case violation of a specification.

  4. 2024

    ICML best paper on model stealing

    In 2024, he is a coauthor of Stealing Part of a Production Language Model, an ICML 2024 Best Paper with Nicholas Carlini, Daniel Paleka and others. The attack recovers the embedding projection layer of black-box production models such as OpenAI's ChatGPT or Google's PaLM-2, and extracts the entire projection matrix of OpenAI's Ada and Babbage models for under $20.

  5. 2024

    CVPR best paper on human feedback

    In 2024, he is a coauthor of Rich Human Feedback for Text-to-Image Generation, one of the CVPR 2024 Best Papers. The paper presents the first rich human feedback dataset for image generation and a multimodal Transformer trained to predict that feedback.

  6. 2024–2025

    AI security at ServiceNow

    He serves as Research Lead for AI Safety and Security at ServiceNow Research from 2024 to 2025 and founds its AI security research team. His research leads to the framework used for AI security testing at ServiceNow.

  7. As of October 2026

    Securing AI agents at DeepMind

    He is a senior staff research scientist at Google DeepMind, where he leads efforts on the development of secure and trustworthy AI agents. He also mentors a Summer 2026 research stream in the MATS program.

Career

  1. Since 2017Google DeepMindResearcher, Co-founder
  2. 2024–2025ServiceNowResearch Lead for AI Safety and Security
  3. 2016–2017Office of ScienceResearcher
  4. California Institute of TechnologyPostdoctoral Fellow
  5. ServiceNow ResearchFounder

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.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

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

  1. Krishnamurthy (Dj) Dvijotham - Research Lead dvij.github.io. Checked 21 October 2026.
  2. Krishnamurthy Dvijotham (Dj) - MATS Mentor matsprogram.org. Checked 21 October 2026.
  3. A Dual Approach to Scalable Verification of Deep Networks arxiv.org. Checked 21 October 2026.
  4. ICML 2024 Awards icml.cc. Checked 21 October 2026.
  5. Stealing Part of a Production Language Model arxiv.org. Checked 21 October 2026.
  6. CVPR 2024 Announces Best Paper Award Winners cvpr.thecvf.com. Checked 21 October 2026.