Google Topics API Privacy Audit
WWW 2026
Measured whether Google's Topics API can be abused to infer users' demographics and sensitive interests from browsing history. Found predictive signals in 17 of 19 traits.
- Python
- scikit-learn
- JavaScript
I'm a final-year PhD candidate at Northeastern University.
My research focuses on building tools to understand security and privacy issues on the web. Lately, I've also been building AI agents that can perform general tasks, and exploring the implications of agent skills. I write about what I find, and I'm actively looking for full-time roles and postdoc positions in industry or academia.
Projects
View all →CCPA Privacy Signal Measurement System
PoPETS 2024(4)
Built a distributed crawler and audit pipeline to measure CCPA opt-out signal propagation across thousands of top websites, revealing widespread failures in enforcement.
Third Parties — Web Almanac 2025
HTTP Archive Web Almanac
Co-authored the Third Parties chapter of the 2025 HTTP Archive Web Almanac, analyzing third-party resource usage across millions of websites.
- BigQuery
- SQL
- Python
Google Search Market Share Analysis
★ Best Paper Honourable MentionICWSM 2024
First large-scale analysis of Google Search's market share under vertical segmentation, using real browsing trace data from a US panel.
- Python
- Apache Spark
- AWS
News
- Mar 2026Industry talk accepted at PEPR 2026 — Why Consent Fails in Practice
- Jan 2026Paper on Google's Topics API accepted at WWW 2026
- Jan 2026Co-authored the Third Parties chapter of the 2025 Web Almanac
- Jun 2025Presented Topics API poster at Google Ads Privacy Conference, Mountain View
- Nov 2024Received travel grant; attended IMC 2024 in Madrid
- Jul 2024Presented PETS 2024 paper in Bristol, UK
- Jun 2024Best Paper Honourable Mention at AAAI ICWSM 2024