Mobile-based academic evaluation systems, such as Computer-Based Tests (CBT), are vulnerable to cheating, including impersonation (proxy test-taking) and illegal collaboration, thereby necessitating more reliable monitoring mechanisms to safeguard exam validity. This study aimed to develop and implement a Face Recognition feature utilizing Google ML Kit as a monitoring system in the CBT Edu application and to examine user responses to its implementation. The research employed a research and development (R&D) method using the ADDIE model (Analyze, Design, Development, Implementation, Evaluation), with exam participants and prospective users as the subjects and Google ML Kit as the research object. Data were collected through questionnaires used to assess system functionality and user experience. The results show that the Face Recognition-based exam monitoring system was successfully designed and implemented in the CBT Edu application. Functional testing using white-box and black-box methods validated all features, while device compatibility and bandwidth consumption testing demonstrated stable performance across various devices and network conditions. User responses measured using the User Experience Questionnaire (UEQ) fell into the “very good” category, with the six dimensions of Attractiveness (5.93), Perspicuity, Efficiency, Dependability, Stimulation, and Novelty (5.72) achieving average scores above 5.7. These findings indicate that the integration of Face Recognition in the CBT Edu application is effective in enhancing exam validity and provides a practical solution to minimize cheating in mobile-based academic evaluation systems, while simultaneously delivering a positive user experience.