Anduril is developing software at the intersection of machine learning, computer vision, and sensor fusion. We apply convolutional deep learning techniques to multi-spectral sensor inputs for autonomous detection and classification. Lattice Ghost UAS, Sentry Tower and other autonomous platforms work in tandem to fuse these detections using Bayesian inference. Our engineers also apply classical techniques in computer vision to perceive environmental features for pose estimation and GPS-denied navigation. Estimation and Tracking engineers at Anduril develop cutting edge software for autonomous operation and situational awareness.