Condition monitoring and fault/life-time prediction for machines and equipment with mechanical parts based on the time-series data of their vibration signals. Our ZETA AIoT solution deploy compressive sensing and embedded machine learning to make sense of the collected data at the edge, so as to allow fast deployment of a durable system that can be added onto existing industrial operations at a very low cost.