Overview

Relling develops multi-modal AI that learns from raw video and sensory data (video, depth, IMU, audio, force, gaze) to better understand and interact with the real world. Its models are trained on high-fidelity measurements to directly capture events, aiming to handle challenges like blur and occlusion in dynamic scenes. The approach uses a broader sensor suite to build 3D perception and robust world understanding, rather than relying on text or static images alone. The goal is to protect critical infrastructure by enabling AI that can reliably perceive, predict, and act in real-world environments.