Our research focuses on building agents that are effective and natural teammates. This involves developing agents (whether robots or software agents) that are not only capable in the world, but also can interact with humans or other agents as part of a team. This research goal draws upon planning and scheduling, machine learning, human-robot interaction, and cognitive science, among other disciplines. Example avenues of work include developing approaches for planning and execution on several different robotic platforms, work on goal reasoning (e.g., goal-directed autonomy), and using techniques from cognitive science to direct learning over long periods of time.