The polynuclear growth model (PNG) is a prototypical example of random interface growth among the Kardar-Parisi-Zhang universality class. In this talk I will discuss a q-deformation of the PNG model recently introduced by Aggarwal-Borodin-Wheeler. We are mainly interested in the large time large deviations of the one-point distribution under narrow-wedge (droplet) initial data, i.e., the rare events that the height function at time t being much larger (upper tail) or much smaller (lower tail) than its expected value. Large deviation principles with speed t and t^2 are established for the upper and lower tails, respectively. The upper tail rate function is computed explicitly and is independent of q. The lower tail rate function is described through a variational problem and shows nontrivial q-dependence. Based on joint work with Matteo Mucciconi and Sayan Das.