SEM24

Given the increase in the number of applications per job offer, companies are finding it increasingly difficult to manage all the CVs they receive, with major issues at stake in terms of time, performance and accuracy. The partner company has developed and marketed a semantic engine (SEM) to match candidates to job offers by extracting relevant information from a CV. The SEM24 project aims to improve the system in terms of accuracy and recall as well as its updateability by drawing on interdisciplinary perspectives. The first aim is to identify relevant new skills and add them automatically to the system using deep learning, from a multilingual perspective. The ultimate aim is to reduce the risk of timely applications being excluded and also to increase the accuracy of the match between candidates and job offers.

To improve the tool's accuracy, the SEM24 project will use hybrid and multilingual methods, combining internal company knowledge with structured and unstructured data, such as feedback from the recruitment process, CVs and job descriptions. In this way, the match between candidates and job offers will be refined to take account of the mastery of skills, as well as social and behavioral competencies. This project will provide essential interdisciplinary input in the field of AI applied to human resources management, and has already been the subject of presentations as part of Master's courses at EHL and at the 9th ‘Human-AI Teaming Workshop’ at the HES-SO Valais Wallis.