Learning an additional language (L2) requires a tremendous amount of effort exerted over an extended period of time. The essence of this effort is captured by the construct of engagement, which refers to the quantity and quality of learners’ active involvement in the process of learning. Engagement is a context-dependent, multidimensional construct consisting of behavioral, cognitive, affective, and social dimensions—each of which is interrelated with and influences the others. Although the topic of engagement has received substantial attention within the broader field of education, studies of engagement in L2 learning are limited. However, given the unique processes which drive L2 learning (e.g., attention, affect, interaction), understanding the factors which may impede or promote engagement is paramount for advancing the field of second language acquisition (SLA). Moreover, considering the context-dependent nature of the construct, understanding how engagement varies within and across instructional contexts (i.e., different skill-based language classes) and scales of activity (i.e., activity at the task, class, program, and sociocultural level) is also needed. Taking an ecologically informed approach, this dissertation study explores students’ and teachers’ perceptions of L2 engagement, as well as factors that visibly impact student engagement across different scales of activity and instructional settings. A case study design was used, with the case defined as an intensive English program at an English-medium university in Thailand. Qualitative data from interviews, post-task reflections, and observational fieldnotes, along with supplementary data from classroom materials and other curricular and policy documents, were collected/generated over one academic term to support the analysis and interpretation of the case. The triangulation of findings from the different data sources (i.e., students, teachers, and classroom observations) indicated that engagement varied across tasks, classes, and engagement subdomains in relation to learning-ecological factors at multiple scales of activity. Students’ and teachers’ perceptions were found to converge and diverge in notable ways, particularly with regard to how members of each category viewed their own roles in effecting engagement. While students’ individual interest in different language skills—a factor initially predicted to impact engagement—was associated with variation in students’ quality of engagement, factors related to the classroom, program, and wider sociocultural context further explained the observed variation. The study contributes a highly contextualized example to the growing body of literature on L2 engagement, which future studies may expand upon in several directions.