Automated systems are increasingly used in recruitment and job applications but, due to proprietary rights, they tend to be opaque and difficult to assess. In this project, you will design an audit of LinkedIn’s automated resume ranking system for job applications. You will develop sets of identical resumes in terms of experience, education and all other job-relevant characteristics where they differ only in controlled characteristics such as inferred gender, nationality or inferred migration/visa status, inferred faith/religion, inferred sexual orientation, or inferred socio-economic background. You will then create one or more job postings and submit the crafted resumes, measuring and quantifying differences in the behaviour of LinkedIn’s automated ranking with respect to these personal characteristics.
School of Computer Science and Statistics