KI på kanten

Offentlig prosjektsammendrag

This project aims to develop a secure, lightweight, and resilient software architecture that enables hybrid AI processing—combining edge inference with asynchronous cloud computation—within the Celerway Edge Foundation platform. The need arises from the limitations of current AI systems, which often assume stable connectivity and rely on heavy cloud integration, making them unfit for mobile or mission-critical environments. Key R&D challenges include ensuring autonomy during network disruptions, maintaining secure and isolated local AI execution, and optimizing for resource-constrained hardware. To solve this, we will design a modular architecture supporting local/cloud orchestration, leverage Celerway’s multipath WAN for robust connectivity, and apply methods such as microservice design, chaos testing, and zero-trust security to validate resilience, efficiency, and interoperability in real-world conditions.