Along with artificial intelligence (A.I.), machine learning is regarded as one of the most in-demand areas for tech employment at the moment. Machine learning engineers develop algorithms and models that can adapt and “learn” from data. As a result, those who thrive in this discipline are generally skilled not only in computer science and programming, but also statistics, data science, deep learning, and problem solving.
According to Burning Glass, which collects and analyzes millions of job postings from across the country, the prospects for machine learning as an employer-desirable skill are quite good, with jobs projected to rise 36.5 percent over the next decade. Moreover, even those with relatively little machine-learning experience can pull down quite a solid median salary:
Dice Insights spoke with Oliver Sulley, director of Edge Tech Headhunters, to figure out how you should prepare, what you’ll be asked during an interview—and what you should say to grab the gig.
What are the challenges faced in a machine learning position?
“You’re going to be faced potentially by bosses who don’t necessarily know what it is that you’re doing, or don’t understand ML and have just been [told] they need to get it in the business,” Sulley said. “They’re being told by the transformation guys that they need to bring it on board.”
As he explained, that means one of the key challenges facing machine learning engineers is determining what technology would be most beneficial to the employer, and being able to work as a cohesive team that may have been put together on very short notice.
“What a lot of companies are looking to do is take data they’ve collected and stored, and try and get them to build some sort of model that helps them predict what they can be doing in the future,” Sulley said. “For example, how to make their stock leaner, or predicting trends that could come up over they year that would change their need for services that they offer.”
How do I prepare for a machine learning interview?
Sulley notes that machine learning engineers are “in rarified air” at the moment—it’s a high-demand position, and lots of companies are eager to show they’ve brought machine learning specialists onboard.
“If they’re confident on their skills, then a lot of the time they have to make sure the role is right for them,” Sulley said. “It’s more about the soft skills that are going to be important.”
Many machine learning engineers are strong on the technical side, but they often have to interact with teams such as operations; as such, they need to be able to translate technical specifics into layman’s terms and express how this data is going to benefit other areas of the company.
“Building those soft skills, and making sure people understand how you will work in a team, is just as important at this moment in time,” Sulley added.
What questions are typically asked during a machine learning interview?
There are quite a few different roles for machine learning engineers, and so it’s likely that all these questions could come up—but it will depend on the position. “We find questions with more practical experience are more common, and therefore will ask questions related to past work and the individual contributions engineers have made,” Sulley said.
- How does Deep Learning differ from Machine Learning?
- What do you understand about selection bias?
- What is the difference between inductive and deductive learning?
- Tell us about a time when you’ve used deep learning methods or machine learning engineering to fix a problem within the company.
- How much of the work did you do yourself?
What are the most important machine learning skills I should know?
A lot of data engineering and machine learning roles involve working with different tech stacks, so it’s hard to nail down a hard and fast set of skills, as much depends on the company you’re interviewing with. (If you’re just starting out with machine learning, here are some resources that could prove useful.)
“For example, if it’s a cloud based-role, a machine learning engineer is going to want to have experience with AWS and Azure; and for languages alone, Python and R are the most important, because that’s what we see more and more in machine learning engineering,” Sulley said. “For deployment, I’d say Docker, but it really depends on the person’s background and what they’re looking to get into.”
What qualities make me a good candidate?
Sulley said ideal machine learning candidates posses a really analytical mind, as well as a passion for thinking about the world in terms of statistics.
“Someone who can connect the dots and has a statistical mind, someone who has a head for numbers and who is interested in that outside of work, rather than someone who just considers it their job and what they do,” he said.
As you can see from the following Burning Glass data, quite a few jobs now ask for machine-learning skills; if not essential, they’re often a “nice to have” for many employers that are thinking ahead.
What should I ask?
Sulley suggests the questions you ask should be all about the technology—it’s about understanding what the companies are looking to build, what their vision is (and your potential contribution to it), and looking to see where your career will grow within that company.
“You want to figure out whether you’ll have a clear progression forward,” he said. “From that, you will understand how much work they’re going to do with you. Find out what they’re really excited about, and that will help you figure out whether you’ll be a valued member of the team. It’s a really exciting space, and they should be excited by the opportunities that come with bringing you onboard.”