Identity verification during in-person evaluations is essential to ensure transparency throughout assessment periods in higher education institutions. This measure is particularly relevant due to potential issues related to identity fraud, which compromise the integrity of academic processes. This article presents the design, implementation, and validation of a low-cost embedded system for real-time student identity verification using facial recognition. The system was developed using an NVIDIA Jetson Nano board, an HD webcam, and an HDMI display, all integrated into a 3D-printed enclosure. At the software level, the system employs open-source libraries face_recognition, dlib and OpenCV within a Linux environment. These tools implement algorithms that encode each face into 128-dimensional feature vectors for comparison using Euclidean distance. The prototype enables student identification in under one second. Experimental tests were conducted using a database of 60 users under controlled conditions, achieving a 100% recognition rate for faces without accessories, and over 95% accuracy under minimal variation. The system proved to be functional, portable, and precise, positioning itself as a viable alternative for enhancing security in academic environments.