Building an End-to-End De-Identification Pipeline for Advertising Activity Data at LinkedIn

Learn how LinkedIn built a comprehensive de-identification pipeline for advertising activity data in this conference talk from PEPR '25. Discover the design and implementation of an end-to-end system that processes billions of events daily while maintaining strict privacy standards and regulatory compliance. Explore the privacy-enhancing technologies applied, including data minimization techniques and privacy-by-design principles that protect user information while enabling valuable analytics for advertisers. Examine the system architecture challenges faced when balancing scalability requirements with business needs, and understand how LinkedIn successfully onboarded performance reporting and billing systems onto this robust pipeline. Gain practical insights into navigating evolving privacy regulations, adapting legacy data pipelines to modern privacy standards, and implementing solutions that satisfy both regulatory requirements and business objectives in the advertising technology space.