Designing an Online Recruitment Platform Based on a Recommendation System

The problem of effective staff recruitment with the desire of utopian “accurate hits”, when both employers and employees fully meet expectations of each other, has always been and will be relevant in the field of recruitment. The initial search for a potential candidate by the employer and the search for a job by the employee implies a time-consuming process of analysis. Of all the many alternatives available on the labor market, seekers need to identify a suitable subset for the consideration. Today, there are several types of online recruitment platforms that aggregate applications of both sides and rank them according to the resume-vacancy matching score offering the most relevant, according to their algorithms, items. However, current solutions are based on the approach of exact search algorithms or manual routine work with big data, rather than on using of modern machine learning algorithms. This study examines an online recruitment platform that adds new features to the candidates’ description by asking them questions that are predefined for their stated skills, questions about their personal attitude to a particular life or work aspect, and, possibly, job-specific questions. On this basis, when candidates profiles have a more detailed feature description and become more comparable, it is possible to build a bidirectional recommender system based on likes and dislikes of a candidate as a whole, one of his or her skills or one of his or her answers (or the lack of answer) to a particular question. The results of this study will cover the first stages in the process of creating software (market analysis, requirements specification and architecture design) and bring the Internet recruitment platform based on the recommender system to the development stage.