Vitaly Shmatikov

Timeline

1. 2000

   A Stanford doctorate

   In 2000 he completes a Ph.D. at Stanford University. The dissertation is Finite-State Analysis of Security Protocols.

2. 2004

   From Moscow to Austin

   He grows up in Moscow and studies as an undergraduate at Moscow University and the University of Washington. Graduate work at Stanford yields an M.S. in engineering-economic systems and a Ph.D. in computer science. In 2004 he joins the University of Texas at Austin as an assistant professor, after work as a computer scientist at SRI International on security, anonymity, and privacy.

3. 2008

   The Netflix Prize dataset

   In 2008, with Arvind Narayanan, he shows how to break the anonymity of the Netflix Prize dataset. That work receives a Test of Time award in 2019.

4. 2015

   Privacy-preserving deep learning

   With Reza Shokri, he publishes Privacy-Preserving Deep Learning in 2015. The paper shows that participants such as hospitals, research institutions, and companies can train neural-network models together by sharing selected model parameters while keeping raw training data on their own systems.

5. 2016

   Joining Cornell Tech

   In 2016 he joins Cornell Tech as a computer science professor. The new professors' work includes machine-learning privacy and security.

6. 2018

   Membership inference attacks

   In 2018 he shares the PET Award with Reza Shokri, Marco Stronati, and Congzheng Song for the 2017 paper Membership Inference Attacks against Machine Learning Models.

7. December 2023

   Text recovered from embeddings

   In December 2023 he shares an Outstanding Paper Award at EMNLP for Text Embeddings Reveal (Almost) As Much As Text, with Alexander Rush, Volodymyr Kuleshov, and Jack Morris. Vec2Text reconstructs 92 percent of a data set of original text exactly, and the team retrieves 94 percent of first names and 95 percent of last names from clinical notes.

8. Since June 2026

   Safety for AI agents

   An Amazon partnership establishes AI4AI, Assured Integrity for AI-Based Software, in June 2026. He is a professor of computer science at Cornell Tech and a principal investigator with Alexandra Silva. The project develops safety protocols for AI agents and the code they produce.