Federated learning (FL) is a distributed training mechanism in which a machine learning model is trained with data that is local to a set of edge devices, e.g., PCs or mobile devices. In this demonstration, we show how we tackle the challenges of implementing a FL system through a combined effort of Qualcomm Technologies, Inc. and Microsoft. We demonstrate this system in two parts. In a technical demonstration for machine learning (ML) researchers we show FL in action on Snapdragon devices as well as training orchestration through Microsoft Florida. Secondly, federated user verification based on voice samples serves as a practically relevant example for the INTERSPEECH community. We feature enrollment, subsequent acceptance of the enrolled user and rejection of impostors using the model that was trained through FL.