We present the Surrogate-gradient Online Error-triggered Learning (SOEL) system for online few-shot learningon neuromorphic processors. The SOEL learning system usesa combination of transfer learning and principles of computa-tional neuroscience and deep learning. We show that partiallytrained deep Spiking Neural Networks (SNNs) implemented onneuromorphic hardware can rapidly adapt online to new classesof data within a domain. SOEL updates trigger when an erroroccurs, enabling faster learning with fewer updates. Using gesturerecognition as a case study, we show SOEL can be used for onlinefew-shot learning of new classes of pre-recorded gesture data andrapid online learning of new gestures from data streamed livefrom a Dynamic Active-pixel Vision Sensor to an Intel Loihineuromorphic research processor.