Matching Performance Metrics with Potential Candidates : A computer automated solution to recruiting

Abstract: Selecting the right candidate for a job can be a challenge. Moreover, there are significant costs associated with recruiting new talent. Thus there is a requirement for precision, accuracy, and neutrality from an organization when hiring a new employee. This thesis project focuses on the restaurant and hotel industry, an industrial sector that has traditionally used a haphazard set of recruiting methods. Unlike large corporations, restaurants cannot afford to hire dedicated recruiters. In addition, the primary medium used to find jobs and job seekers in this industry often obscure comparisons between relevant positions. The complex infrastructure of this industry requires a place where both recruiter and job seeker can access a standardized overview of the entire labor market. Introducing automation in hiring aims to better address these complex demands and is becoming a common practice throughout other industries, especially with the help of internet based recruitment and pre-selection of candidates. These solutions also have the potential to minimize risks of human bias when screening candidates. This thesis aims to minimize inefficiencies and errors associated with the existing manual recruitment screening process by addressing two main issues: the rate at which applicants can be screened and the quality of the resulting matches. This thesis first discusses and analyzes related work in automated recruitment in order to propose a refined solution suitable for the target area. This solution –semantic matching of jobs and candidates - is subsequently evaluated and tested in partnership with Cheffle, a service industry networking company. The thesis concludes with suggestions for potential improvements to Cheffle´s current system and details the viability of recruiting with the assistance of an automated semantic matching application.