Peter Kairouz
Research Scientist at Google
Palo Alto, California, United States
Summary
👤
Senior🎓
Top SchoolPeter Kairouz is a research scientist at Google with eight years of experience focused on decentralized, privacy-preserving, and robust machine learning at scale. His work bridges theory and production: contributions to high-profile open-source projects like TensorFlow Privacy and TensorFlow Federated include implementing DP queries, RDP computations for heterogeneous subsampled Gaussian mechanisms, and stateful DP aggregation utilities. Trained in electrical engineering and applied mathematics, he brings deep expertise in differential privacy, distributed algorithms, and wireless/communications from a trajectory spanning academia (Stanford, UIUC) and industry research. Based in Palo Alto, he combines rigorous theoretical grounding with practical engineering that enables privacy guarantees in real-world federated learning systems. An underappreciated thread in his career is how his signal-processing and communications background informs efficient, noise-aware privacy mechanisms across decentralized systems.
9 years of coding experience
6 years of employment as a software developer
University of Illinois Urbana-ChampaignBachelor of Engineering in Electrical and Computer Engineering, Image and Signal Processing, Bachelor of Engineering in Electrical and Computer Engineering, Image and Signal Processing at American University of BeirutPostdoctoral Research Fellow, Data Privacy and Artificial Intelligence, Postdoctoral Research Fellow, Data Privacy and Artificial Intelligence at Stanford University
Arabic, English, French
Github Skills (12)
machine-learning10
privacy10
differential-privacy10
tensorflow10
federated-learning10
python10
testing9
data-structure7
data-structures7
algorithm7
algorithms7
data-analysis6
Programming languages (1)
Github contributions (5)
tensorflow/privacy
Jan 2019 - Feb 2022
Library for training machine learning models with privacy for training data
Role in this project:
ML Engineer
ML EngineerContributions:5 commits, 14 comments, 1 issue in 3 years 1 month
Contributions summary:Peter primarily contributed to the development of privacy-preserving machine learning techniques within the TensorFlow Privacy library. They implemented queries for sum and average calculations, including weighted variants. The user also added functionality to compute RDP for heterogeneous applications of the subsampled Gaussian mechanism and integrated a Distributed Skellam Query for DP.
machine-learning-modelstraining-datamachine-learningprivacy
google-parfait/tensorflow-federated
Sep 2019 - Mar 2022
An open-source framework for machine learning and other computations on decentralized data.
Role in this project:ML Engineer
ML EngineerContributions:14 commits, 6 comments in 2 years 6 months
Contributions summary:Peter implemented utilities and tests for integrating differential privacy within the TensorFlow Federated framework. Their contributions included developing a `build_dp_aggregate` function that supports stateful aggregation for differentially private queries, demonstrating expertise in privacy-preserving machine learning techniques. The user also developed the tests for these utilities. The commits show a focus on ensuring the correct implementation of privacy-related functionalities within the federated learning context.
machine-learning
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.