Designing for data privacy
summary
Company migrates scattered PII into a single auditable vault using tokenization with Piano and an abstraction layer, replacing phone-number-as-ID and removing PII from data warehouse and logs. A 33 million–record French breach motivates the change; rollout starts with one payment microservice, exposes issues — inconsistent team implementations, blanket detokenization, and cultural resistance — and forces practices like deterministic vs pointer tokens for uniqueness and updates. Outcome: deletable records, masked front ends, analytics on tokens only, and a product principle privacy over practicality.
AI-written summary · about
chapters (9)
- 0:00Data privacy challenges
- 2:34Unregulated to GDPR compliance
- 5:47Privacy architecture options
- 7:52Tokenization design and microservices
- 9:40Tokens reduce breach risks
- 11:47Rolling out tokenization
- 13:55Overcoming team resistance
- 15:00Privacy over practicality
- 16:46Current state and zero trust
related talks
16mWhat if privacy had an API?
50mPrivacy by design
01h 59mDeep Dive Workshop – Data Security and Data Protection
01h 59mDeep Dive Workshop – Data Security and Data Protection
28mHow Privacy Engineers Deliver ROI
26mAnalytics & Data Science with Privacy at its Core
12mStriking the Balance: Safeguarding Customer Privacy While Empowering Employees
16mPrivacy by Design for JavaScript Data Systems