SensiML is an automated software tool that allows IoT and wearable device developers to easily build embedded firmware that transforms raw complex sensor data into meaningful pattern matching events. Through a process of collecting example data within the SensiML software, the tool produces SoC-specific optimized code running directly on the class of low-power microcontrollers most used within sensor endpoints themselves.
SensiML-generated algorithms complement cloud analytics for demanding applications, allowing sensors to pre-process insight events in real-time rather than stream large volumes of raw sensor data for high latency cloud processing. The tool's simplicity belies its sophistication with advanced AI employed behind the scenes to enable application domain experts of modest programming skill to create highly accurate sensor algorithms without need of coding or data science expertise. The product of four years of development within Intel as the Knowledge Builder Toolkit, the core team acquired the technology and formed SensiML as a hardware-agnostic developer solution.
SensiML now supports the Arm Cortex-M processor family with its toolkit, delivering the ability to generate powerful ML algorithms with edge sensor inference and learning capability.