Hiring from the Gild
A computer science or IT degree from a reputable university looks great to employers hiring in the big data space. But how well does it serve as a predictive indicator?
Gild, a company devoted to optimizing Human Resources and recruiting initiatives, is trying to answer that question through compiling and analyzing big data.
In a world where prowess with a computer is equally, if not more, important than social skills, glowing recommendations from college professors and top professionals may be less important than advertised. Further, since data science is a relatively new profession, universities may not do as good a job of preparing prospects for the field as in other more established fields.
“The traditional markers people use for hiring can be wrong, profoundly wrong,” said Dr. Vivienne Ming, Gild’s chief scientist. Here, Dr. Ming is referring not only to the world of programming as a whole but also to one of Gild’s own top programmers in Jade Dominguez.
Gild found Dominguez using an algorithm that scoured the web for programming talent using metrics like the amount of times a person’s code was reused by other programmers. Places like GitHub that provide an open source platform for programmers to share and improve upon their code have facilitated the collection and aggregation of these metrics.
For example, Gild discovered that Dominguez’s code had been reused 1,267 times. That metric combined with others spat out a score of 100 on Gild’s fledgling talent-finding algorithm. “We did our own internal gold strike,” Dr. Ming said. “We found this kid in Los Angeles just kicking around his computer.”
Another portion to Gild’s hypothesis is that intuition and name recognition influences hiring more than it should. Data analytics, according to Ming, should play a bigger role in predicting a person’s prospective performance than it is.
Determining data science prowess can be rough, but Gild hypothesizes that sample sizes in IT are indeed big enough for data driven insights to be meaningful. As evidenced by both this site’s big data job bank and Tableau’s big data job openings visualization covered here last week, there exist thousands of opportunities in the big data space. Likewise, there exists thousands of data science and architecture jobs being done right now and companies presumably keep information on both performance and background.
Again, Gild’s concept is in its infancy. One major hurdle to clear is that of determining a coder’s usefulness. The company’s algorithm that measures other programmers’ respect for the coder is a good start. More research and an expansion of the relevant datasets could further Gild’s presence in IT hiring. Until then, Gild’s idea mostly remains just that—an idea, albeit an interesting one.