All news

Face value: Russian IT company's new algorithm identifies person’s occupation by photo

The algorithm also defines a candidate’s wage expectations

MOSCOW, March 29. /TASS/. SuperJob, a Russian HR portal, has created a neural network enabling headhunters to identify an applicant’s occupation by photo. A representative with the company told TASS the accuracy of making a resume via the algorithm is 88%.

"So far, the neural network can already easily match anyone with one of 500 basic professions. For example, there is 99% probability that the system can tell a driver from a salesman or an environmental engineer based on a photo," the representative said.
He added that there is a 98% probability that the algorithm can define whether an applicant has any higher education or not, as well as his or her age and work experience. The algorithm also defines a candidate’s wage expectations. The network has the base of 10 mln clothing samples.

"We know how much these clothes cost. It is not without a reason that the saying "clothes make the man" emerged. By observing how a candidate is dressed on the photo, the system tallies up his or her wage expectations, head of the company Alexei Zakharov said as quoted by the press service.

According to the algorithm developers, this internet recruiting technology became possible after the analysis of more than 25 mln photos from CVs. They note the more a photo suits the occupation of a candidate, the easier it will be to make up a resume. Besides that, when the algorithm creates a resume, the applicant will be able to correct it himself.