We introduce WaLRUS, a wavelet-based SSM leveraging SaFARi for improved accuracy and stability in modeling non-smooth, transient signals, outperforming traditional HiPPO-based models.