C-Elect, An Automatic Candidate Selection Tool Using Constraint Programming

Whether it is for training, recruitment, or admission to a school, it is common to implement a call for applications process and then carry out a pre-selection of candidates. How-ever, the task can become challenging and risky, mainly when dealing with an extensive list of candidates to be evaluated or a plethora of criteria that need to be considered. In addition, this selection process may be subject to inequities, potentially leading to unfairness. Nowadays it is usual to automate the selection process efficiently and optimally. Several automated selection approaches are designed to handle this problem, but most rely on machine learning models that require historical data to train their algorithm. This paper introduces C-elect, a generic system based on constraint programming, which is not dependent on historical data. Subsequently, an online platform is proposed to automate the candidate selection process, specifically focusing on the pre-selection stage. C-elect enables a user to carry out an automatic candidate selection operation by uploading: a compressed file of candidates' CVs, an Excel or csv file containing information on the various characteristics of the candidates. The system then dynamically generates a constraint-based programming model appropriate to the data entered and the criteria specified. Extensive experimental tests were conducted to evaluate its performance and effectiveness. The results were promising. The system demonstrated its capability to efficiently sift through a large pool of candidates and generate shortlists of individuals who met the user's predefined criteria. This automation saved a significant amount of time and ensured that candidates were selected based on objective and consistent evaluations.