Published June 25, 2020 by Scott Hirsch
This is the last of five posts about machine learning at TalentMarketplace. Take a quick read of the introduction blog here.
Nicholas, a PhD student in Condensed Matter and Materials Physics at Simon Fraser University, joined Applied Quantitative Methods as part of his passion project. Through AQM, Nicholas set off to create an intricate machine learning process for TalentMarketplace for a smooth employer experience.
When employers search for candidates, they may be overwhelmed at the selection of candidates available, unsure which candidate to pursue first. As such, Nicholas was tasked to come up with a solution to ensure a time efficient and accurate process for employers by narrowing down the selection of quality candidates.
Thus, came the Candidate Interview Prioritization Tool. With this tool, candidates are ranked based on their likelihood of successfully completing an interview, in which scores are based off of previous profiles passing interviews and the interviewer’s past decisions. A score between 0 and 1 is assigned to each profile, in which higher scores indicate a higher likelihood of passing the interview, helping employers know where to spend their time.
By using the data points captured on TalentMarketplace, a machine learning model has been trained using historical data to learn skill combinations that have resulted in a successful or unsuccessful interview to enter the TalentMarketplace talent pool.
What struck as interesting about this tool are the number of potential users of it. Not only could this tool be applied to our clients who are interviewing multiple candidates, it could also be used by TalentMarketplace internally for ranking candidates who apply to our platform in the event we see a large influx of new candidate users. Further, I have wondered if this could be productized and offered to our clients - or even our competitors - as a standalone product which can be utilized to optimize their interviewing processes.