Several seismic techniques, both passive and active, exist for estimating the shear-wave velocity Vs structure of shallow sedimentary structures. In particular, passive compliance signals recorded by broadband ocean-bottom seismometers (OBSs) can be used to invert for Vs structure. While compliance-based imaging studies have been carried out at several locations across the Cascadia Subduction Zone, such an approach has not been extensively applied to OBSs deployed on the continental shelf and slope. In this study, we measure compliance and coherence signals at 13 broadband OBSs deployed along Cascadia’s continental shelf and upper slope. We then use a recently developed technique to probabilistically invert compliance signals for shallow Vs structure that makes use of mixture density neural networks. Finally, we compare and contrast our inverted Vs profiles and derived properties obtained using this method with previous studies focused on the properties of basin sediments.