Elie Bursztein

Elie Bursztein leads Google's anti-abuse research team, which invents transformative solutions that help protect users against online threats.

Recent

blurry image for loading
blurry image for loading
cryptography
A Hacker’s guide to reducing side-channel attack surfaces using deep-learning
A Hacker’s guide to reducing side-channel attack surfaces using deep-learning
Talk Defcon & BH 2020
This talk showcases SCALD, our tool that leverages deep-learning explainability and dynamic execution to automatically find which parts of a crypto-hardware implementation is responsible for leaking the information exploited by side-channel attacks
blurry image for loading
blurry image for loading
security
Malicious Documents Emerging Trends: A Gmail Perspective
Malicious Documents Emerging Trends: A Gmail Perspective
Talk RSA 2020
This talk provides a comprehensive analysis of the malicious documents that target users and corporate inboxes, an in-depth analysis of the latest evasion tactics used by attackers and what Google is doing about it.
blurry image for loading
blurry image for loading
web
Five years of the Right to Be Forgotten
Five years of the Right to Be Forgotten
Publication CCS 2019
We conducted a retrospective measurement study of 3.2 million URLs that were requested for delisting from Google Search over five years.
blurry image for loading
blurry image for loading
security
Protecting accounts from credential stuffing with password breach alerting
Protecting accounts from credential stuffing with password breach alerting
Publication Usenix Security 2019
In this paper, we propose a privacy-preserving protocol whereby a client can query a centralized breach repository to determine whether a specific username and password combination is publicly exposed, but without revealing the information queried.
blurry image for loading
blurry image for loading
cryptography
A Hacker Guide To Deep Learning Based Side Channel Attacks
A Hacker Guide To Deep Learning Based Side Channel Attacks
Talk Defcon 2019
This talk provides a step-by-step introduction on how to use deep learning to perform AES side-channel attacks.
blurry image for loading
blurry image for loading
anti-abuse
Deconstructing the Phishing Campaigns that Target Gmail Users
Deconstructing the Phishing Campaigns that Target Gmail Users
Talk Black Hat USA 2019
In this talk we look into Gmail telemetry to illuminate the differences between phishing groups in terms of tactics and targets.
blurry image for loading
blurry image for loading
ai
Rethinking the detection of child sexual abuse imagery on the Internet
Rethinking the detection of child sexual abuse imagery on the Internet
Publication WWW 2019
In order to scale CSAI protections moving forward, we discuss techniques for automating detection and response by using recent advancements in machine learning.
blurry image for loading
blurry image for loading
user experience
Understanding the online safety and privacy challenges faced by South Asian women
Understanding the online safety and privacy challenges faced by South Asian...
Blog post Jun 2019
For South Asian women, a major hurdle to their meaningful participation online is their ability to ensure their safety. This post illustrates this challenge by recounting the safety and privacy challenges faced by women across India, Pakistan, and Bangladesh, who talked to us about their online experiences.
blurry image for loading
blurry image for loading
ai
Cutting Edge TensorFlow - Keras Tuner: hypertuning for humans
Cutting Edge TensorFlow - Keras Tuner: hypertuning for humans
Talk IO 2019
In this talk, we will introduce Keras Tuner, a hypertuning framework made for humans.
blurry image for loading
blurry image for loading
user experience
They Don't Leave Us Alone Anywhere We Go - Gender and Digital Abuse in South Asia
They Don't Leave Us Alone Anywhere We Go - Gender and Digital Abuse in South...
Publication CHI 2019
To better understand the gendered risks and coping practices online in South Asia, we present a qualitative study of the online abuse experiences and coping practices of 199 people who identified as women and 6 NGO staff from India, Pakistan, and Bangladesh, using a feminist analysis.