Designing an optimal team is of interest in many disciplines: recruitment of faculty members to build an effective academic department, matchmaking in online video games to create fair and interesting experiences, and building skillful and efficient teams in stock trading firms. The difficulty inherent to this problem is that teams cannot be fully understood by studying their members in isolation as team performance is not simply a sum of individual performances. Moreover, it is not yet understood how the patterns of interactions and relationships among team members impact performance. In this regard, we need to build upon theories from sociology, computer science, and dynamical control to study team behavior. To this end, I plan to study how individual attributes and the communication patterns between individuals impact team performance. I also plan to study a data-driven empirical model inspired from structural balance theory—a classical theory describing the dynamics that govern the sentiment of interpersonal relationships—and assess its impact on performance. I plan to use ideas from multi-task learning and reinforcement learning to predict and optimize teams' performance and design rational sequential decision making policies. I will conclude with my future research directions.