SAN FRANCISCO — The nation's best jobs boast salaries that average $100,000 and up, offer generous company benefits, and promise to have recruiting suitors fighting for your hand.
But they are highly technical roles carrying job descriptions like DevOps engineer and analytics manager that demand an alphabet soup of computer skills as well as incessant on-the-job learning.
So do you have to be a math genius with a spare PhD in physics to get one of these great gigs? What we found might surprise you.
USA TODAY conducted a series of interviews with people in the nation's top jobs, as well as with those recruiting and training candidates for these roles. Our mission was to see just how obtainable these plum posts are, whether you're a student looking for a profession or a career-changer seeking better pay in a booming field.
The market for tech talent is so hot that technology jobs now rate as the best type of employment opportunities in the nation.
A 2017 survey of the top jobs in America by employer ranking and assessment site Glassdoor — based on earning potential, job satisfaction and number of openings — ranked data scientist, DevOps engineer, data engineer, and analytics manager as the 1, 2, 3, and 5 top jobs (the fourth was tax manager). That's also the second year in a row data scientist took the top slot.
Other jobs surveys, including one from search site Indeed.com, have echoed these results, reflecting a cloud computing-driven surge in information capture.
The good news? Not only do you not necessarily have to be a rocket scientist, although we did speak with one, some folks in these roles are entirely self-taught.
But you'll need free time to acquire or polish skills and languages (Python, SQL, AWS to name a few), plus often money for training sessions. While some we talked to honed their skills solo, often using online courses, many recommended an intense type of trade school called a boot camp. Some are inexpensive, but others can cost five figures.
You also don't need to be a numbers whiz — or even like math. But distilling massive quantities of information requires comfort with computing and data sets.
“One problem our industry has is there’s really no school for it,” says Leslie Carr, an infrastructure engineering manager and former senior DevOps engineer with Clover Health, which uses data science to improve healthcare for seniors. “Not only is there no degree in DevOps, if someone did have a degree from four years ago and hasn’t being doing it all, their knowledge would be out of touch.”
That all may seem a bit daunting. But the flip side is that these jobs are plentiful and are expected to grow.
They also aren't just at tech companies in Silicon Valley, Austin or Seattle, but may very well be at a healthcare startup in Nashville, an auto manufacturer in Detroit, or just about any company that wants to analyze their data to improve the bottom line.
Companies are eager to fill these positions because “it’s not just software that’s eating the world, but data, because it informs every decision companies now make,” Nidhi Gupta, senior vice president of engineering at Hired.com.
What you need to know
The specific skills required for these jobs is a mix of programming languages and ability to communicate data findings in simple language.
Here’s a quick dive into the job descriptions (according to Glassdoor) as well as three critical skills associated with each (provided to USA TODAY by two tech recruiting firms, CyberCoders and ZipRecruiter):
Data scientist: Analyze raw data sets to extract learnings and insights. Skills needed: Machine Learning, Python, R, SQL (according to CyberCoders); Statistics, Machine Learning, Python (ZipRecruiter). Median base salary: $110,000 a year.
DevOps engineer: Familiar with IT infrastructure software development in order to work across teams. Skills needed: Linux, Amazon Web Services, Chef (CyberCoders); Linux, AWS, Puppet (ZipRecruiter). Median base salary: $110,000.
Data engineer: Work closely with data architects and scientists to prepare data so it can be analyzed. Skills needed: Hadoop, Python, SQL (CyberCoders); SQL, ETL, Computer Science (ZipRecruiter). Median base salary: $106,000.
Analytics manager: Responsible for determining strategies for collecting data, implementing solutions and conducting research. Skills needed: Google analytics, data analysis, marketing spent metrics (CyberCoders); Communications skills, statistics, SQL (ZipRecruiter). Median base salary: $112,000.
“Five to 10 years ago, the tech jobs were all about web and mobile development, which was largely about coding,” says Ryan Aylward, chief technology officer at Glassdoor. “To be successful now in these data-focused roles, you have to have an analytical mindset," says Aylward. "With the right foundation, you can be OK with a variety of backgrounds.”
Plus, despite much talk of late about automation slowly killing traditional jobs, data science is to some degree robot-proof as it demands a human touch to creatively solve problems with the help of computers.
How to get them
From a self-taught actress and a military medic to a baseball-crazed statistician and plasma physics scholar, the backgrounds of people who have found their way to data science jobs is broad. But wherever they started, they had to pick up the specific skills these jobs demand.
Some universities are tweaking their curriculums to acknowledge this skill-set shift. But where colleges fall short, other training organizations have jumped into the gap. Some are either free or low cost and specifically focus on supporting women, minorities and veterans, such as Code.org, Code2040, Operation Code and Vets Who Code.
Paid bootcamps up the ante and the promises. These include Galvanize, whose six-month data engineering course costs $17,000 and offers a solid track record of employment for graduates.
“The supply imbalance of open jobs and the lack of skills is pretty staggering,” says Galvanize CEO Jim Deters, who likes to call his venture “the people’s MIT.”
Deters says his students have included Best Buy clerks, frozen-yogurt stand workers and a homeless person. “We’re looking for somebody that has the inclination and the right attitude towards learning, and we’ll give them the skills,” he says.
At the University of Washington, the place where Bill Gates honed his computer lab skills, students increasingly are being exposed to the concept of managing tremendous amounts of data, says Bill Howe, associate director of the school's eScience Institute.
“We’re trying to find a set of requirements that different departments can meet in different ways,” says Howe, citing analytics, data management and visualization as key focus areas. “These days, being good at data means you could get a job anywhere. Nordstrom, the post office, Walmart, they’ve all got analytics groups.”
Who they are
If there is a commonality between these data-science tech workers, it's both a passion for figuring out what huge sets of numbers mean and a skill for distilling those conclusions in a way that ultimately helps a business, be it a shoe retailer or a media company.
Clover Health's Carr majored in chemical engineering in college, and then switched to a tech support role that eventually landed her gigs at Google and Twitter.
Carr says degrees don’t matter to her when assessing talent. Instead she looks for Python and Ruby programming skills, as well as a job candidate’s contributions to open source code, a computing standard that is open to all and encourages users to build on the discoveries of peers.
George Xing, an analytics manager at ride-hailing company Lyft, took his Princeton degree in economics and computer science and went to work on Wall Street as a fixed income analyst.
When he moved to San Francisco in 2012 looking for a change of pace, “no one was doing analytics, but since then it’s become a standard function to solve business problems using data.”
Xing says data-science job recruits at Lyft “have to have math skills and tech skills, as well as the ability to communicate what their findings mean in simple language to non-technical people.”
'I hate math,' says one DevOps pro
But not everyone goes the collegiate route. Alex Spence, whose armed forces resume includes a stint as a medic in Iraq, taught himself information technology skills while in high school. Now he handles DevOps roles as lead product engineer at social media startup Sprinklr in Austin.
“I’ve never taken calculus, the furthest I got was geometry and algebra 2,” he says. “I hate math.”
Daphne Dorman, who handles DevOps for medical tech company Vineti (formerly Vitruvian Networks), is totally self-taught, a former actress who now volunteers her time helping other members of the transgender community break into tech roles.
“You get one life in this time, in this body, so if you’re interested in something you should just go for it,” says Dorman, who describes her tasks as maintaining servers, handling security issues and keeping code up to date. “It’s all about the infrastructure that allows engineers to do what they do.”
Dorman says the key to success in data science is “a strong desire to be here and to contribute,” noting that especially at small startups it’s impossible to hide behind fancy degrees. “Without a deep curiosity and desire to constantly learn, it’s tough to last,” she says.
Theresa Johnson, a data scientist with Airbnb, would seem like someone who could coast through any tech job. She earned a bachelor’s, master’s and Ph.D from Stanford University, specializing in aeronautics and astronautics with a focus on plasma physics. So yes, a rocket scientist.
And yet when Johnson decided she wanted to pivot to tech, she felt she “still had a lot to learn.” So she buckled down with a few online classes with Coursera to see how her skills set would need to be adapted to data science.
Now, when she isn’t sorting through huge data sets to help Airbnb’s website deliver better lodging matches faster for consumers, part of her role is helping fill job openings in her new field.
Who is she looking for to join the data science revolution? The skilled, sure. But mainly the passionate.
“People looking to get into these jobs need to come with a familiarity with statistics, enjoy math and science, and love to test hypotheses,” she says. "This isn’t about building things that are automated. This is all about finding out, ‘Why does this data matter?’ You have to dig in.”