Protecting Sensitive Data in Web Apps
summary
Shows a concrete pipeline: encrypt bytes on the device with client-side encryption, store ciphertext in MongoDB, and process only inside a hardware-isolated secure enclave that auto-provisions keys so operators cannot read plaintext. Demo code is JavaScript, uses MobileNet (TensorFlow Lite) deployed into the enclave, and the commands cape run and deployment encrypt the function and data. Demo encrypts ~1,100 mouse points and classifies uploaded images (panda, Egyptian cat, Grace Hopper → 92% military uniform), and the flow works end-to-end.
AI-written summary · about
chapters (4)
- 0:00Start
- 2:23Keeping Data Protected is Hard
- 5:04Encryption is Complicated
- 8:51Encryption for Developers
related talks
21mProtecting Sensitive Data and Machine Learning Models
40mProtecting Sensitive Data and Models for Machine Learning
34mEncrypting Data in the Browser - Exploring Web Crypto APIs
39mEncrypting Data in the Browser - Exploring Web Crypto APIs
23mKeeping secrets with JavaScript
05mEnd to End Encryption for Mobile Developers
42mYour Data is Showing
24mStrictly Confidential: Use Cases & Demos of AI Workflows in TEEs