Deploying intelligent systems at the edge presents a unique set of challenges. These decentralized architectures operate with limited resources and bandwidth, making them vulnerable to a range of threats. From adversarial access to data violations, safeguarding edge AI deployments requires a layered