Five tips to find and hire data scientists
- 03 December, 2012 10:40
- Comments
Data scientist jobs are hot in the era of big data. These professionals are responsible for manipulating petabytes of data into better decision-making, new streams of revenue and ultimately more business. A study by McKinsey Global Institute shows that a company using big data to its full potential could increase its operating margin by more than 60 percent, but there's a big shortage of talent.
CIOs are struggling to find and hire people with that perfect balance of business acumen, database expertise and effective communication skills.
"Everybody wants these people and it's harder to find them," says John Reed, senior executive director at Robert Half Technology, an IT staffing firm in the US. "Because the demand far outweighs the supply, you will have to go to more sources for the right candidates."
You won't likely find many candidates the phrase "data scientist" on their resume. Some candidates may not even know they are a fit for the data scientist job. Here are five tips for finding and hiring a data scientist.
1. Look for a team instead of one person
You may find candidates with some skills and not others. So you'll have to warm up to the idea that you may not find everything in one person, which means hiring two or more people to fulfill your needs.
According to an EMC survey of 497 data scientists and business intelligence workers, half of big data scientists partner frequently with other data scientists, statisticians or programmers.
Tom Soderstrom, CTO of NASA's Jet Propulsion Laboratory (JPL), says that although he can clearly define the data scientist role at his company, he knows it may be impossible to find. "It's a special type of person and I've discovered that I don't think they exist." Instead, he says, "It could be a team of people. A data scientist could work with several interns and a community around them."
At NASA JPL, Soderstrom sees the potential for data scientists to manipulate satellite data about oceans or weather to develop new types of science experiments. He says he's looking for someone who can both speak the language of business and work with big data technologies, such as Hadoop. "It's someone who can teach data to tell an interesting story that we didn't already know."
2. Look internally
Soderstrom is undertaking a thorough search for external candidates, but he's also aware of the possibility of grooming internal candidates -- an option that is usually cheaper and has the added benefit that the person would be familiar with the JPL's environment and customers.
Looking at mid-level engineers and programmers, Soderstrom says he is "shaking the trees to see what comes out." He plans to hire a data scientist this month and surround that person with interns to help with some of the legwork. He also wants to carve out a career track for interns to become data scientists.
Reed recommends investigating whether there are employees who have the raw analytics skills and would benefit from an investment in training. He says the only downside is that internal candidates still have to learn as they go and may not be as proficient at certain tasks.
3. Sell them on your company
If your company decides to fish in the pool of external candidates, Reed says it's important to have a strong and compelling story to tell because the up-and-coming tech firms and hot startups will be tough competitors for talent.
"You have a cool story to tell if you work for Google or Facebook, but what if you don't?" he says. "You have to get them excited about working [for your company]. You are selling them on joining your organisation. Unless you offer compensation and benefits and have a great story to tell, these people will [only] want to work on the coolest projects out there."
4. Look for consulting skills
JPL's Soderstrom says that one of the qualities he looks for in candidates is a consultant mindset. "A good consultant goes in and talks to a customer with a problem," he says. "He understands the problem and comes up with a creative solution with the customer."
If you're unable to find a person with that mindset, using an external consultant may be your next best bet. Reed suggests enlisting one if the idea of growing your own data scientists isn't appealing or won't work for your company. But he warns against hiring someone who "may have the academic side of it, but not the practical work experience."
5. Join networking and technical user groups
To find candidates for a data scientist job, you need to find out where the candidates spend time. Reed recommends making yourself and your company visible. "You have to make sure these people can find you where they go to network." Reed suggests joining LinkedIn groups, social media groups and technical user groups. "Go on an awareness campaign," he says.
Until universities produce more graduates with big data skills to address the shortage, CIOs will need to be creative to find people with the right skills and the right fit for their big data aspirations.
Contact Lauren Brousell at lbrousell@cio.com and follow her on Twitter at @LBrousell.
Recommended
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.
- Bookmark this page
- Share this article
- Got more on this story? Email CIO
- Follow CIO on twitter
-
Samsung Galaxy S4 vs. HTC One: 5 Reasons to Choose the GS4
-
Samsung Galaxy S4 vs. HTC One: 5 Reasons to Choose the GS4
-
Solving the skills conundrum – part 1
-
Samsung Galaxy S4 vs. HTC One: 5 Reasons to Choose the GS4
-
High school students still see ICT as ‘sitting at a computer all day’: survey
-
Detecting APT Activity with Network Traffic Analysis
Today’s successful targeted attacks use a combination of social engineering, malware, and backdoor activities. This research paper will discuss how advanced detection techniques can be used to identify malware command-and control (C&C) communications related to these attacks, illustrating how even the most high-profile and successful attacks of the past few years could have been discovered. -
Top 10 Reasons You Don’t Need MDM
Even though companies such as Fiberlink make it easy to centralise the management and security of mobile devices of all shapes and sizes, that is no excuse for capitulating to the whims of the masses, to bolster your career by trying to be hip and trendy and au courant. If you allow progress here, next thing you know, the company will be asking you to enable business in the cloud. Where will it all end? So, before you take that fatal next step in allowing mobility, security, BYOD and user flexibility, carefully read these top 10 reasons why you don’t need mobile device management. -
Benefits of Deploying Microsoft Exchange Server 2010 on Dell Compellent with Data Progression
Messaging and collaboration platforms have emerged as mission critical applications, consuming a large portion of IT spending for organisations. The rich features in these applications have significantly changed the messaging requirements and needs of today’s information from anywhere with any device, the result is an ever increasing demand on storage systems both in terms of capacity and bandwidth. Many organisations are rethinking their storage strategies to meet the demanding criteria and to handle the future requirements. Read more.
Get exclusive access to Invitation only events CIO, reports & analysis. 
