Based on our team’s deep expertise in areas like fraud detection, risk-modeling, and streaming time-series ML prediction problems, we have invented technology that allows us to rapidly learn custom event classes from new sources of streaming time-series multi-sensor data. These events are not just simple events or anomalies but can extract multi-level semantic structure from your data and customize themselves using sophisticated ML algorithms. They can also scale and deploy on either an AWS stack or your own on-premises stack.