#privacyengineering

What Should an LLM Be Allowed to Remember? LLM memory can make financial applications feel...

A media site should make consent a first-class authorization record, then place the CAPTCHA decision...

NotebookBloom: Shipping analytics without collecting user content I needed a real activation funnel...

The MS Paint invisible watermark embeds a server GUID into every AI image — even when inference runs locally. Here's how to detect it.

A self-hosted Python scanner that detects trackers, fingerprinting scripts, missing security headers, and leaky cookies on any website. Free, fast, and runs locally.

Every prompt you send to an AI gets logged somewhere. I built a local audit middleware that encrypts, stores, and lets you search your own AI conversations without trusting anyone else's server.

Local inference is the easy half of on-device AI without breaking E2EE. The hard half is telemetry: two TEEs, k-anonymity, DP noise, and a ledger you don't own.

Why Privacy Is the Most Misunderstood Concept in Data Science A technical follow-up to...

Healthcare research and responsible AI both depend on access to useful data. But in practice, many...

WTF is this: Differential Privacy Edition Ah, the joys of trying to understand the latest tech...