Nicolas Papernot
Old Toronto, Ontario, Canada
Summary
🤩
RockstarNicolas Papernot is an Associate Professor at the University of Toronto and a Canada CIFAR AI Chair at the Vector Institute whose work sits at the intersection of security, privacy, and machine learning. He blends rigorous academic research—recognized by an Alfred P. Sloan Research Fellowship—with hands-on contributions to influential open-source projects like TensorFlow models, TensorFlow Privacy, and the CleverHans adversarial example library. His technical output spans differential privacy, private learning with ensembles of teachers, machine unlearning, and adversarial robustness, and he translates theory into practical tutorials and tooling used by the community. Beyond academia he maintains ties to industry through ongoing collaborations with Google Brain, and his research has attracted mainstream press attention, reflecting both technical depth and real-world relevance.
11 years of coding experience
Stackoverflow
351reputation
8kreached
24answers
0questions
Github Skills (22)
adversarial-attacks10
python10
machine-learning10
mask-rcnn10
keras10
differential-privacy10
tensorflow10
faster-rcnn10
dcgan9
model-driven9
cgan9
privacy9
modeling9
model-driven-development9
data-analysis9
Programming languages (7)
Github contributions (5)
cleverhans-lab/cleverhans
Sep 2016 - May 2020
An adversarial example library for constructing attacks, building defenses, and benchmarking both
Role in this project:
ML Engineer
ML EngineerContributions:4 releases, 5 reviews, 927 commits in 3 years 8 months
Contributions summary:Nicolas appears to be involved in the development and refinement of adversarial example techniques within the CleverHans library, which focuses on building attacks, defenses, and benchmarking machine learning security. Their contributions include fixing import issues, adding support for specifying the epsilon value in the MNIST tutorial's FGSM attack and incorporating the inclusion of parameter passing to the Fast Gradient Method. Several commits also include code changes related to the Saliency Map Method (JSMA) and associated tutorial improvements, including bug fixes and removing of the model call dependency on Keras learning phase.
benchmarkingmachine-learningsecurity
tensorflow/privacy
Dec 2018 - Nov 2019
Library for training machine learning models with privacy for training data
Role in this project:ML Engineer
ML EngineerContributions:2 releases, 67 commits, 5 PRs in 11 months
Contributions summary:Nicolas primarily contributed to the development of the MNIST tutorial within the TensorFlow Privacy library, including modifications to the loss function and kernel initializers. They enhanced the tutorial's functionality by adding ReLU activations. Furthermore, the user added support for Eager mode and Keras to enable training with DP-SGD.
machine-learning-modelstraining-datamachine-learningprivacy
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.