Dawn Hudson - Fotolia

Hiring software developers? You need AI

It's a great time to be a developer, but an awful time to try to hire one. Hiring software developers requires skill, patience, luck, and, just possibly, artificial intelligence.

With demand wildly outstripping supply, hiring software developers that fit might feel akin to winning the lottery. Actually, winning the lottery might be easier.

But that is all about to change. Advances in artificial intelligence (AI) and cognitive computing may give employers a new tool to make hiring software developers easier. And some of those same technologies might also make it easier to decide who to keep -- and who to fire -- in a merger and acquisition situation.

Why do these technology advances matter? Because nearly every company in the world is, or will be, involved in designing and releasing software, and that demand has led to a severe and worsening worldwide shortage of developers.

"Without doubt it's a software developer's market," said Andrew Heyes, managing director of Harvey Nash Southern Regions, Harvey Nash UK, a well-known worldwide recruitment firm. "In the 2016 Harvey Nash Technology Survey, we asked almost 3,000 technology people how often they were headhunted. Software developers came [out on] top with more than six in 10 having had at least 10 qualified approaches in the last year."

So, as an employer, how do you stand out to potential prospects?

That's where source{d}, based in Madrid, Spain, comes in. Founder and CEO Eiso Kant, who has been coding since he was 12, saw time and again at different jobs the difficulty of hiring software developers. It didn't make sense to reach out to them just based on skill sets, he thought, because there's much more to the job than that. When, with his latest startup, Tyba, "We couldn't find developers no matter how many eyeballs we had on job posts," Kant explained he and his team realized it was time to rethink the entire process. He wanted a way to cut through the noise, leave the built-in biases behind and look at not just what developers know, but how they work.

Using what Kant called "AI learning and deep learning," his team built an engine that can reach out to big repositories of code like GitHub, BitBucket and Savannah and index and download every project. "We've been able to analyze over 5 million developers so far. It's amazing." And, with GitHub adding an estimated 200,000 new developers sharing code monthly, there will always be more places to look. "We analyze all 5 million developers across all version control of the code so we can see the latest snapshot. This represents 1 PB of data, and, if you laid all the code out end to end, it would go around the world," he said.

But it doesn't stop there. "I used to get really upset about how some recruitment people really weren't putting developers first," Kant added. So once the AI engine looks at the data, source{d} uses live developers with real-world experience to reach out to candidates with jobs that might be a good fit for them. "We always have a developer reaching out to a developer," he said, with the goal of figuring out strengths, weaknesses, interests and motivations. Currently, source{d} reaches three developers a minute via email. If developers aren't interested, the team makes sure to find out what might tempt them to move and then stores that data to be used when the right job comes along. "In the end, we've built a company of developers, for developers," Kant said.

But say you're in the position of having a few too many people and needing to decide who stays and who goes. Cognitive computing may come to the rescue here as well. Another startup, Broad Listening, harnesses the power of a "cognitive computing system and a social emotional simulator" to sort through email, logs, team interactions and other data to compile 100 personnel metrics on existing or potential employees, CEO and Founder Sarah Austin said. "We want to be able to detect all the social, emotional and levels in humans," she said, and gather and sort all that data easily and quickly. This system can pick up a lot of different "unbiased cues" about people based on what is said in casual email exchanges -- everything from gender or age bias to time management issues, "imposter syndrome," or those abusing the system.

Currently, Broad Listening is targeted at the human resource space, but Austin said her dream is to broaden the underlying cognitive computing technology and make it much more of a learning and training resource for employees. "We could create a coach or a personal assistant that just pulls information from all the siloed data and makes it easy for employees to get the best advice on a wide variety of things."

Next Steps

Looking for developers? Surprisingly, it's not just a paycheck that matters

Given up? Try outsourcing

Cognitive computing's secret weapon? Watson

Meet your new colleague: The machine software tester

Dig Deeper on Topics Archive