This paper considers the problem of regulating a nonlinear dynamical system whose state cannot be directly measured. Leveraging neural networks to approximate system states, we demonstrate input to state stability of the closed-loop of the dynamical system and our data-driven controller.