Control-Oriented Neural State-Space Models for State-Feedback Linearization and Pole Placement