710650Machine Learning for Computing Occupational Culture-fit using Automated Predictive Video Recruitment SoftwareClosedGRD Proof of ConceptInnovate UKLaunchPad Recruits is a Software as a Service business who provide an online video interviewing platform. Set up in 2012, LaunchPad Recruits have quickly established themselves as a major player in the video interviewing market by implementing proprietary software with reviewing and comparison functionality. Success has come from adapting candidate screening requirements to customer needs and providing a superior candidate experience. The power of video interviewing is in part due to the wealth of potential data from the candidates’ presentation, personality and body language, which future software realisations seek to exploit. This project aims to take video interviewing to a completely new level, by developing an automated predictive video recruitment (APVR) application. The proposed APVR application will predict which candidates out of a large pool of applicants meet the performance and cultural fit requirements for a particular role. Interviewer bias and inconsistencies between different human reviewers are removed, to make sure that the top talent is identified. This revolutionary screening process is made possible by bringing together a number of cuttingedge technologies such as machine learning, computational personality profiling and predictive analytics. In addition to the established benefits of video recruitment, namely reducing travel costs and shortening the timescales to fill positions, the APVR application can save organisations the cost of recruiting poor performers. One study found that 46% of new hires fail within the first 18 months, and for most this was due to attitude and not lack of skills to do the job. The cost to employers in managing a bad hire can be four to six times the employee’s salary. Therefore the proposed APVR application has the potential to become an extremely valuable recruitment tool, also applicable in sectors where screening of large candidate pools is required.