Task selection is a critical element of mathematics teaching because mathematical tasks differ in the mathematical opportunities made available to students. Thus, it is important to examine both the tasks that PSTs chose and why they were chosen. This study examines the task selection of 10 pre-service teachers (PSTs) in a middle grades mathematics methods course. Each week, PSTs prepared and delivered 90-minute lessons for their assigned small group of middle grades students in an after-school enrichment program a local middle school. PSTs were free to choose the content of their lesson plans. I use Remillard’s (2005) framework of the teacher-curriculum relationship paired with a documental approach to didactics to infer PSTs’ instructional aims and their personal and pedagogical resources leveraged during task selection and lesson planning. PSTs’ lesson plans, lesson reflections, and semi-structured interviews were analyzed to identify the intellectual resources, perspectives, and epistemologies employed by PSTs when preparing their lessons. Three broad instructional aims shared by PSTs are identified. For each of the three themes, I describe the shared aim and demonstrate how it combines with other personal resources to form a scheme of utilization which informs PSTs’ participation with instructional resources. First, the enrichment sessions should be fun. PSTs differed in how they conceptualized fun, attending to either the structure or the mathematics of the tasks. Second, PSTs aimed to avoid surprises during their lessons by anticipating student responses. Finally, PSTs aimed to select or create tasks that “fit” their students. PSTs assessed task fit by the absence of unproductive struggle and whether students completed the task. This study identifies several productive beliefs and dispositions held by PSTs when selecting tasks in an early field experience. These beliefs and dispositions can be leveraged by teacher educators to support the development of ambitious teaching practices. Furthermore, this study demonstrates the importance of modeling high cognitive demand tasks in both mathematics methods and content courses.