How to Hire a Remote Machine Learning Engineer

Pamela Bump

Pamela is Managing Editor at Emerj. She previous worked in B2B digital publishing with Innovation Leaders and Boston MedTech. Pamela holds a Master's degree in Media Ventures from Boston University.

How to Hire a Machine Learning Engineer

When searching Indeed at the time of writing this article, over 770 remote machine learning job listings were posted. A search on LinkedIn yielded over 1,200 results.

While the remote hiring theme seems to be a growing trend in the areas of machine learning, it may also be a trend throughout a variety of other job positions as well.

Of the business leaders who attended last year’s Global Leadership Summit in London, 34% said more than half of their full-time staff would be working remotely by 2020. Furthermore, a recent survey from the US Bureau of Labor reported that one-third of the country’s workforce has a part-time, contract, or freelance position other than their full-time job.

In the tech industry, Toptal notes that companies like GitHub and Dell have taken on mass remote hirings around the world. While Dell, like the above-noted business leaders, says over half of its employees will work remote by 2020, Toptal also reports that one third of GitHub’s employees work out of office.

More and more developers, especially AI experts, are looking to work remotely or in more flexible arrangements. This trend means that smart, innovative companies may to be considering this pool of labor in addition to hiring for onsite staff. Otherwise, they may miss out or face tough competition from enterprises such as Google, which can outbid for local talent.

We spoke to Toptal’s Director of Engineering and AI Lead, Pedro Alves Nogueira, to get his insights on how an employer should hire and onboard a remote machine learning expert. He also spoke about hiring processes that could work across the board for lead AI experts and their future team members.

Before Toptal, Nogueira received a PhD in Artificial Intelligence and Human-Computer Interaction from the University of Porto in Portugal. He graduated summa cum laude. Prior to this, he served as a user experience and human-computer interaction researcher at several European and international projects, including on the Carnegie Mellon – Portugal Programme.

Identifying Problems and Data

As a company prepares to enter the process of hiring a machine learning expert, they must first look at the data they have, as well as the data they might need for the new employee to create working machine learning models.

“Having a good data infrastructure and making sure you actually have data for this person to work with is important,” Nogueira says. “It’s one of the milestones you need to reach before you actually start thinking about hiring someone.”

While big companies may have a plethora of data, he added that smaller to medium-sized companies may still be building datasets.

In the latter scenario, “hiring someone before you start collecting data might be advantageous because this person might look at your business, what you would be able to achieve and what would have the highest impact. It might actually help steer these efforts which would actually save you money.”

Whether they are helping to build up data or creating new models, Nogueira says the first machine learning hire will often be the person who leads projects and has a hand in hiring their future team.

Because of this, he says it’s important for at least one leader from the company’s engineering side and one stakeholder from the less-technical business side to meet and determine technical needs, business goals, and other qualifications needed for the machine learning role.

According to Nogueira, the two types of stakeholders who are most likely to make the best hiring judgments could be:

  • One tech-minded person who understands the data and has the strongest possible statistical or mathematical education: “Someone who understands the technical models and the aspects of putting them together. This includes production, getting access to the data, determining what type of access and what type of data. or determining processing power to crunch the numbers.”
  • One business-side stakeholder, such as a leader in sales or marketing, who’s going to see a “benefit” from the candidates’ models

Nogueira says the best way to start the hiring process is to have the above-noted team members draft a job description together.

While he says having an AI engineer speak for the technology team’s needs is ideal, he notes that companies can also build solid job listings with guidance from a current software engineer who has a hard science, technological, or statistical background.

He adds that it is important that both sides clearly understand what goals are expected from this role so they can make the best-educated guess:

Make sure [both parties] actually understand the data that they have, the problem they’re facing, what the expectations are. Then they post that as a job posting or give it to us as a description so that we can actually find the right individuals.

Attracting Talent

While Nogueira did not speak on the process of attracting talent, we have seen from our research that companies have gotten creative with this process.

In January 2018, the MIT Technology review noted that major technology companies attended December’s Neural Information Processing Systems (NIPS) conference where they held parties and gave out swag specifically to attract the various AI engineers that attended. Because this was so recent, it was unclear if these strategies reported any further results.

Aside from offering swag, research from Atlas suggests that a remote work life may be its own perk within the technology field. When Atlas asked programmers to mark a list of their preferred job benefits, vacation or days off was preferred by over 57%, while remote work was the next highest preferred benefit at 53%. With this in mind, businesses may want to start looking for freelance talent when remote work is applicable. According to the study, a majority of software developers and machine learning engineers seem to prefer the flexibility of remote work, and so offering it to them might give businesses in need of their talent an advantage.

The chart below shows other perks that survey respondents marked:

Atlas's survey on programmer preferences
Atlas’s survey on programmer preferences

Hiring: Build a Bridge First

As a company is beginning to hire for AI or machine learning efforts, Nogueira says the goal is to find someone who is a strong, articulate communicator that understands both the technology and business needs:

I call these people the bridges because they actually make the bridge between the technical and the non-technical parts. These people have to wear multiple hats, because essentially, I see them as the elite AI engineer.

This person has the abilities of an AI engineer, but would also be able to lead or build a team that’s technical and can communicate with people who are not technical,” he explains. “They need to educate people, and they need to communicate what they are able to do and not able to do.”

In finding a bridge person, the engineering and business stakeholders who created the job listing should be involved in the interview and selection process, according to Nogueira.

While an already-hired AI engineer or a software engineer with strong mathematical or statistical knowledge will know what technical questions to ask, Nogueira says the business-oriented person will be able to speak with the candidate about company goals.

“Have this person talk with the candidate so that they can understand the business, the needs, and the requirements,” he notes.

In interviews for any type of AI talent, Nogueira says two of his favorite questions are:

  • What are some of your biggest achievements?
  • What were the challenges you faced when reaching them?

Nogueira says the candidates’ responses will prove if they’ve actually built models, while also showing how well they can communicate and articulate what they’ve done to both technical and non-technical people.

“You want to make sure that they’re not just someone who knows how to use some online library or API to build some ad hoc models, but then doesn’t know how to create or think about them,” he says.

Make sure that this person isn’t just buzzword-friendly. A machine learning project is not a straight arrow. You have to take this data and these features and build this model. Then when you’re done, you have the best possible model and you’re going to be able to know that it is. You’re going to try out lots of things, you’re going to make assumptions, and you’re going to make different models. There’s a lot of thinking involved and not so much coding in mind.

He adds, “You also want to know that they can communicate and articulate what they’ve done and what they will do even when speaking to non-technical stakeholders.”

While business leaders may want their bridge person to have expertise in the machine learning field and in academia, he says they may occasionally find other individuals that are suitable as machine learning team members.

Some very brilliant people, some of the brightest individuals that I’ve worked with in the past, were not from computer science backgrounds. They did not have a PhD in AI. They were actually a mathematician or statistician coming from physics or hard science. They understood the math. For them, it was actually easier than some developers to get into the math.

He adds, “If I were non-technical, I would look at someone who has a hard science background. Someone who ideally has some track records from existing projects. But most of all, someone who knows how to communicate.”

Consider Dynamic Duos

While some will consider hiring one person who’s an expert in two or more AI technologies, Nogueira says hiring multiple people, each with expertise in at least one strategy, can provide some benefits.

“If you’re just hiring one or two individuals, you’re probably looking for multiple technical trades, or a ‘Master of Some.’” he says. “You’re going to want to find someone who’s exceptionally good at being able to communicate and being able to dive deep into whatever problems you have. This person is going to do it all. … If you’re hiring a team, then you have more options.”

In an example Nogueira gave, he noted that instead of hiring a machine learning engineer with expertise in natural language process and computer vision, an employer could choose to hire one computer vision expert with some NLP experience, as well as one NLP expert with some computer vision experience.

“Then you’re going to put these two individuals together, and what you’re going to see is that these people are inquisitive. They’re going to start talking to learn from one another. They’re going to become thoroughly proficient at what they were not a few months ago. Then, they’re going to have great ideas that they’re going to share with one another,“ he says.

Onboarding: Connecting the Wires

Whether a business leader is hiring a bridge person or a team member, Nogueira still encourages putting them through a proper onboarding process, especially if they are remote.

First, he says, “Make sure people meet. These positions can come with a lack of visibility, so you need to formally introduce people. They need to be shown where they can ask for help.”

Nogueira also adds that the employee shouldn’t just be “dropped” into a project. An employer should make sure that they have access to the data and contacts that they will need.

Additionally, Nogueira explains that when they match a remote machine learning expert with a company, they make sure they contact of the client company hiring them has full access to the data the remote worker may need. They also make sure that the contact at the client company is available to onboard them properly.

This is where that whole integration, flexibility and flow between teams is important. It’s important to have good data readily in real time to be able to massage and make processes and resources to process new incoming data as quickly as possible.

If following the above hiring and onboarding instructions, Nogueira says that if an employer can feel comfortable letting a remote machine learning engineer get to work:

Just connect the wires and give it a jump start. If you have the right people they’ll just do their thing.

Nogueira says he hasn’t seen any failure cases related to onboarding AI engineers because:

Most of the people that we have are very inquisitive. They have been doing machine learning for some time now, even as freelancers. They know how to deal with clients. They know how to communicate really well. They don’t really wait for the client to tell them what to do. They just get access to the data and go into it pretty quickly. … Having someone with that attitude mitigates a lot of the problems, especially in enterprise clients who tend to be big and sometimes bureaucratic.

Use Case: GitHub

One example of a company that may keep remote workers connected throughout onboarding and their time at the company is GitHub. While over one third of GitHub’s international employees are remote, the Harvard Business Review reports that all are required to do one week of orientation at the company’s San Francisco headquarters.

Once employees are remote, HBR says the company offers a “#toasts” forum on its “online water cooler” platform where employees and remote leaders can interact, send photos and give credit to remote employee achievements.

Closing Thoughts on Hiring a Remote Machine Learning Engineer

While hiring a bridge person or other remote engineer can be a daunting task, Nogueira says his strategies have helped him to hire inquisitive leaders who go on to build strong remote teams of their own.

Through our discussion with him, as well as our own research, we’ve noted the following takeaways for business leaders:

  • The hiring process, from determining employee needs to holding interviews should be managed by both engineering leaders and non-technical business stakeholders
  • The ability to strongly communicate past AI-related achievements is key to finding someone who will bridge the gap when communicating with engineers and stakeholders in the future
  • If hiring one person, business leaders should find a Master of Some, but if hiring two people, they can find two experts in related things that will learn from each other.
  • When onboarding, a remote engineer should be shown who to turn to for help and introduced to key teammates, even though employers should look for someone who can autonomously access data to do their job.

As the machine learning talent pool continues to gain more competition, a business leader may want to consider remote hiring for these technological positions.

A recent study from Element AI used LinkedIn profiles and conference survey data to estimate that there are 22,000 PhD-educated AI researchers. Of that 22,000, the study notes that only 3,000 said they were looking for a job, or did not have a current position marked on their LinkedIn profile.

While the study also notes that only half of the researchers were located in the US, business leaders may also want to consider hiring remote in order to access a portion of the talent pool that wasn’t originally geographically accessible to them.

 

This article was sponsored by Toptal, and was written, edited and published in alignment with our transparent Emerj sponsored content guidelines. Learn more about reaching our AI-focused executive audience on our Emerj advertising page.

Header Image Credit: Dice Insights

Subscribe
subscribe-image
Stay Ahead of the Machine Learning Curve

Join over 20,000 AI-focused business leaders and receive our latest AI research and trends delivered weekly.

Thanks for subscribing to the Emerj "AI Advantage" newsletter, check your email inbox for confirmation.