<![CDATA[Employee attendance is an important aspect in village office personnel administration because it is related to discipline, attendance data management, and the preparation of administrative reports. However, the conventional attendance system at the Lingkok Berenge Village Office causes the re-capitulation process to be slow, prone to recording errors, and still allows for the practice of leaving attendance. This study aims to design, implement, and evaluate a face recognition-based employee attendance system to improve identification accuracy and support the digitalization of attendance administration. The developed system, SIMADES-FACE, was built using CodeIgniter 3, PHP, and MySQL and applies a face descriptor-based face recognition method using face-api.js run-ning on TensorFlow.js, with the process of face detection, feature extraction, and matching de-scriptors to biometric data stored in the database. The study used a Research and Development (R&D) method with a quantitative descriptive approach. System evaluation was carried out through black-box testing, measuring facial recognition accuracy, False Acceptance Rate (FAR), False Rejection Rate (FRR), and detection time. The test results showed an accuracy rate of 93.33% from 180 trials, a FAR value of 0%, an FRR of 6.67%, and an average detection time of 2.26 seconds. These results indicate that SIMADES-FACE is able to support the employee attendance process more accurately, efficiently, and digitally documented.]]>
