FireFli is a mobile perception system that integrates complementary sensing modalities to enable robust, continuous monitoring across both line-of-sight and non-line-of-sight environments. The platform constructs a dynamic scene representation of its environment through context inference and multi-object tracking, facilitated by autonomous navigation. By incorporating WiFi-based radio frequency sensing, the system extends its perceptual range beyond physical occlusions, enabling the detection of anomalous human states, including fall detection, without requiring direct visual access to the monitored subject. Together, these modalities produce a context-aware architecture that supports real-time anomaly detection and continuous behavioral monitoring without human intervention. This capability is particularly relevant in clinical and long-term care settings, such as acute care hospitals and geriatric facilities, where there is increasing demand for unobtrusive, scalable monitoring infrastructure that minimizes disruption to patients while maintaining safety oversight.