Recruiting’s Move to Machine-learning Rolls On

Oct 4, 2016
This article is part of a series called News & Trends.

The latest of the machine-learning launches that include the Siri-for recruiting and Watson is Restless Bandit, from the maker of another highly successful recruiting startup.

One of the people behind Restless Bandit, Steve Goodman, was the CEO of Bright, which sold to LinkedIn for $120 million. He says Bright was about pre-screening resumes before they came into an applicant tracking system. Restless Bandit, on the other hand, is a bolt-on to a system to look at candidates who already applied to a job and are in the ATS.  “The best talent is hiding in plain sight,” Goodman says, “and you don’t even know it.”

The company notes that “companies with more than 1,000 employees typically have more than 50,000 resumes in their talent pool. Companies with more than 5,000 employees often have more than 1 million resumes on hand.”

Restless Bandit is a trained algorithm — machine learning. So if hiring managers are frequently saying they like certain people, the machine learns that and the algorithm then gets smarter. It’s matching resumes against a job description. It looks at things like where you worked before (Pepsi will give you a better ranking for a Coke job than if you’re working at John Deere now); career arc (title ascension); skills, though it’s not the most important signal; resume gaps; education and certifications; salary expectations based on their current job; and hundreds of other variables.

Goodman and his team have been working on this new product for about two years, and say they have “aggregated more than 120 million job descriptions and 30 million resumes to train its algorithms.”

They’ve raised $10 million, and gotten IHOP, Applebee’s, Gannett, among others, using it so far.

This article is part of a series called News & Trends.