N.E.G.U

# PredictIQ — Multimodal Machine Intelligence PredictIQ is an IoT-based predictive-maintenance platform that connects physical machine telemetry with Meta AI to make machine conditions easier to understand and investigate. Machines generate important signals such as temperature, vibration, current and RPM. PredictIQ collects these readings, stores machine history and uses its existing deterministic health and prediction engine to identify abnormal operating conditions. During the Meta Global AI Developer Hackathon, I extended PredictIQ with a Meta AI intelligence layer. The system retrieves verified machine readings, health states, alerts and existing prediction evidence, then uses Meta models to generate understandable explanations and assist technicians in investigating machine conditions. The system is designed to distinguish measured facts, existing engine outputs and AI-generated interpretations. AI recommendations do not replace the deterministic machine-health engine. The project explores how multimodal and language models can make industrial machine intelligence more accessible, especially for technicians working with physical equipment. **Theme:** Push the Frontier.