In academic settings, especially in densely populated areas, supervisory roles face multifaceted challenges. Managing communication, discussions, task assignments, attendance tracking, and schedule coordination across varied platforms can be particularly burdensome, also detracting from students’ experiences. Existing mobile applications generally lack the necessary integration for these settings, as they are predominantly designed with industrial project management needs in mind. Our study introduces “Supervision”, a comprehensive mobile application that leverages deep learning and context-aware computing to address these challenges. Apart from integrating the latest mobile application development and intuitive UI/UX technologies, we conducted a comparative analysis of popular facial recognition models (FaceNet, OpenFace, DeepFace, and VGGFace) using the LFW dataset to deploy the most ideal model in real-time face identification on resource-limited devices. Evaluating several criteria (accuracy, inference time, and memory utilization), our findings revealed that FaceNet outperformed other models with 94.04% accuracy, the shortest inference time (0.018 sec), and the lowest memory usage. This study also enhances the efficiency of the supervisory process, promoting timely project completion and fostering robust communication between supervisors and students.