RyzenShield: The Zero-Trust AI Edge Bodyguard | AMD Slingshot Hackathon

Описание: We expose too much to the cloud. From pasting API keys into ChatGPT to uploading unredacted student IDs, every digital action is a privacy leak waiting to happen.

Enter RyzenShield—a flawless, 100% offline edge computing architecture powered strictly by the local AMD Ryzen NPU. We flipped the traditional security model upside down to prove that absolute privacy doesn't have to compromise on blazing-fast speed.

In this demo, we showcase our core Ryzen NPU-accelerated modules:

🛡️ Secure Browser: Real-time interception and "Shadow Token" sanitization of sensitive data before it hits the network.
🧠 Threat Sandbox: A local LLM 'Teach-Back' that dissects phishing tactics and generates isolated Burner Personas.
🖼️ Visual Shield & Document Shield: Local ONNX computer vision pipelines that permanently burn out biometric and PII data from images and PDFs.
🔐 PixelShield Vault: A zero-latency LSB cryptography engine that mathematically camouflages your deepest secrets into the noise of a benign image.

By eliminating the 70.22ms cloud roundtrip, the AMD XDNA engine physically processes complex intelligence locally in under 9 milliseconds. When privacy is processed on the edge, security is undeniable.

Built with power, speed, and zero-trust for the AMD Slingshot Hackathon.

👨‍💻 Created By: • Tharun N V • Kesavan G

🔗 GitHub Repository: https://github.com/TharunvenkateshN/RyzenS...

#AMD #RyzenAI #EdgeComputing #Cybersecurity #ONNX #MachineLearning #ZeroTrust