Hadoop Jobs: How One Exec Vets Talent

16 Top Big Data Analytics Platforms

Ask Think Big Analytics CEO Ron Bodkin what he looks for when vetting job candidates, and you might notice his first few answers appear to have little to do with Hadoop and other big data technologies. Rather, they focus on people.

"It's more art than science," Bodkin said of evaluating big data talent. In an interview, we asked Bodkin, among other things, what he looks and listens for when evaluating candidates for big data-related roles. Here's what he had to say.

Prioritize a person's potential over his past
"Our basic philosophy is to find the overachievers who are super capable and have the right attitude," Bodkin said.

To do so, Bodkin and his team place more emphasis on so-called soft skills than on technology chops. Sure, if you want to be a top-notch data engineer, you're going to need to develop premium programming skills. Similarly, if you want to become a data scientist, a strong background in math and statistics wouldn't hurt. But the underlying skills are more tools for the job than predictors of success in Bodkin's eyes.

"...people [who] have proven they can learn new things, they can drive results, they work well on teams, they collaborate -- those are some of the key qualities that we look for," Bodkin said.

[How connected will future vehicles be? Read 5 Ways Big Data Can Improve Your Car.]

There will probably always be a seat at the table for the brilliant engineer more comfortable with machines than people; there will be many more seats reserved for those who pair sought-after IT skills with the ability to communicate and work well with others.

Figure 1:

"We need to have people that are able to work with our customers... and communicate effectively and understand what they need," Bodkin said. The same ability would hold true with internal partners and stakeholders, too. "Almost anybody building a successful big data team [needs those people]."

Tech skills still matter
People skills are big on Bodkin's list, for sure, and he believes that if you hire good people with the right aptitudes, you can train them on the nitty-gritty of Hadoop, Cassandra, Pig, HBase, Yarn, and other big data technologies. "We teach a lot of technical skills based on hiring the people with the right potential," he said.

Still, you can't ignore tech skills entirely. That CPA might be a whiz with numbers, but that doesn't make her a data scientist. Here are some of the things that Bodkin said he likes to see in a candidate's background, especially in combination:

Next Page

That last one's worth some extra emphasis. Java skills are no doubt a good thing for big data professionals given that Hadoop and related technologies are based on Java. But Bodkin said it's a good sign when people know other languages as well; among other things, it shows a willingness to learn.

"We actually prefer people that don't just have experience in the Java world... they know SQL, they've probably been successful in something else like Python or R or Ruby," Bodkin said.

Don't be too quick to rule people out
If you subscribe to Bodkin's strategy of emphasizing people and their potential rather than specific skills and experience, you'll likely buy in here, too. Although you might want to have your list of red flags handy when reading resumes or meeting with candidates, don't be too quick to judge.

An example: Although Java experience is certainly a major plus for Hadoop and other big data jobs, it's not always a must if a person nails your other criteria -- provided you're willing to train and be patient. When Bodkin was VP of engineering at Quantcast, the firm hired several people who lacked a Java background even though their jobs demanded Java skills, and they ultimately worked out just fine.

"They were able to learn the Java ecosystem; it just took a little longer," Bodkin. "It depends on whether you're able to nurture top talent through a longer learning curve to be successful."

Shore up software engineering fundamentals
Particularly when it comes to senior-level data engineering positions, Bodkin said it's important to remember the basics of successful software engineering: "People who know the discipline and practice of testing, designing effectively, using continuous integration, designing for testability -- [look for] some of those types of skills as well."

Bodkin also sees an emerging trend linking big data -- especially Hadoop-based endeavors -- and the DevOps approach. "Having a bunch of engineering skills but also administration skills in Linux is increasingly important," he said. "We think there's a connection between big data and effective DevOps," and subsequently a growing need for talent in this arena.

Although Bodkin doesn't have a long list of red flags when it comes to identifying and vetting big data talent -- common sense and internal needs should rule -- one that does get his radar chirping: Resumes or LinkedIn profiles overloaded with big data buzzwords, without the requisite evidence of real knowledge and experience in the technologies and responsibilities behind the buzzwords.

"If you see a long list of things [and] it's unlikely that they've had that level of exposure, that's important," Bodkin. "That's certainly something that we like to get some validation of... is this person representing what they're capable of accurately? Do they know what they know and know what they don't know?"

You can use distributed databases without putting your company's crown jewels at risk. Here's how. Also in the Data Scatter issue of InformationWeek: A wild-card team member with a different skill set can help provide an outside perspective that might turn big data into business innovation. (Free registration required.)