AEGIS-X

Aegis-x by Nestlink Team(nestlink.dev)

AEGIS-X is an autonomous, AI-powered cybersecurity platform that detects and neutralizes digital deception in real-time. Built on Amazon Nova's multimodal reasoning capabilities and secured by blockchain verification, it creates an immutable trust layer for digital content across social media, financial systems, and enterprise platforms. It detects deepfake media and flags them for users. The engine is built on top of 3 deep learning models(image swin transformer, video swin transformer and wav2vec for audio files) the engine utilizes 4 agentic layers operating in async mode;Aegis- V (analyzes frames using tools driven by nova 2 lite ,Aegis R is the orchestrator for all the agents,Aegis-M runs media through required deep learning model and Aegis B handle cryptography and and blockchain anchoring on etherium network, this architecture is built on top of fastapi, postgresql, docker,tanstack start etc where we collaborated as a team(Comfortine Siwende Olunga, Melvins Simon, Sheldon Billy, Peter Magana). We were generally inspired by the following massive financial losses; 1.$25 million USD stolen in a single deepfake video conference scam (Arup, Hong Kong 2025)

2.AUD $25.8 million lost to AI voice cloning scams in just six months (Australia 2025)

3.130+ million views of a single political deepfake targeting US elections (2024)

4.₹3.75 crore ($450K) lost to celebrity deepfake investment scams (India 2025)

5.93% of security professionals report increased AI-generated fraud (2025 industry data)