If you’re a tech startup in the AI space, you’re going to need to hire a team of technical experts to develop your product. For many, this will include sourcing top notch artificial intelligence, machine learning, and deep learning engineers. Demand for these technical experts is growing rapidly. The United States cited a 344 percent growth rate for machine learning engineers in 2019. Overall, computer and information technology jobs are predicted to grow by 22 percent in the next ten years, well above the nation’s average.

Despite this exponential growth in the field, hiring tech talent isn’t straightforward. We’ve seen startups make numerous hiring mistakes over and over again, leading to stalled projects, slow growth, and decreased profits. These are the three biggest mistakes we see startups make when hiring AI, ML, and deep learning engineers — and how to avoid them.

The 3 Biggest Mistakes Startups Make When Hiring AI, Machine Learning and Deep Learning Engineers

  1. Not sourcing talent globally.
  2. Hiring based on credentials alone.
  3. Not testing programming skills.

Read More from Built In’s Expert Contributors NetworkArtificial Intelligence vs. Machine Learning vs. Deep Learning: What’s the Difference?


1. Not Sourcing Talent Globally

Despite high demand, there is currently a talent shortage for engineers with machine learning experience. In the United States, especially in tech hot spots like the California Bay Area, big name tech companies like Google and Microsoft tend to hire most of the available local talent, making it difficult for smaller startups to hire as competitively.

To compete, startups need to change their perspective: What if we thought globally, instead of locally, about recruitment?

Today’s remote work environment makes globally sourced talent ripe for consideration. Covid-19 shifted many tech company employees to permanent work-from-home assignments. Not only is remote work more productive, but allowing it opens your company up to a top tier of global talent you may not be able to access otherwise.

Moreover, in some locations worldwide, deep learning engineers may not have the job opportunities accessible in the United States, despite their advanced technical skill. These prospects may jump to work at a startup with an interesting angle or problem to solve, bringing substantial value to your team. 


2. Hiring Based on Credentials Alone

Many companies today automatically filter out job applicants before recruiting staff even look at a resume. Applicants are rejected based on higher education requirements, university name, years of experience, and more. Because of this, it’s no wonder that 50 percent of applicants lie on their resumes.

And, in the fields of AI and ML, a Ph.D. from Stanford isn’t always the best predictor of future performance. In fact, it’s often not.

Why not? Doctoral students are trained to research a problem, publish their findings, and repeat. There is very little technical application to real world problems. In the startup world, you don’t actually need your employees to conduct most of the research in-house. Instead, you need someone who can read academic papers, understand the concepts, derive relevant insights, and apply them to the project that they are working on. If you hire an applicant without applied technical skills, you may quickly regret your decision.

Also consider:

  • Preference for team versus solo work. Building a product is a much more collaborative process than most would think. Make sure your potential candidates work well with a team approach. 

  • Appetite for continuous learning. You’ll need someone to stay on top of the latest trends and research, which is constantly evolving. A candidate set in their ways or comfort zone, who isn’t open to adopting new approaches, isn’t going to be someone you want on your team. 

Read More on Built In’s Expert Contributors NetworkWant a Career in Machine Learning? Here’s What You Need to Know.


3. Not Testing Programming Skills

You’ve expanded your recruitment globally and are vetting applicants for applied skills experience. What’s next? Testing those skills.

While most AI, ML, and deep learning engineers should have the theoretical knowledge you need, not all of them are good programmers as well. If you’re going to ship a competitive product to market, and ship it quickly, you need engineers who can also program well.

You wouldn’t hire a new copywriter without testing their writing skills, would you? This same school of thought should be standard for ML engineer applicants, as well. Often, startups interviewing ML, deep learning, or AI talent focus the interview on theoretical concepts and never test the candidates’ actual coding abilities.

The tests don’t have to be complex. For example, you could assign a candidate a simple research paper and ask them to build the neural network outlined using an open source machine learning platform like PyTorch or TensorFlow. This is a great way to (A) see how quickly they can work and (B) see how they can apply research concepts in a real world scenario.

Better Hires, Better Products

The bottom line is that if you invest quality thought and time in the recruitment process, you’re going to end up with a more marketable, competitive product. This will ensure you build a strong technical team that can both understand cutting edge research and apply new concepts, helping you build a foundation for long-term success in the competitive startup marketplace.

Expert Contributors

Built In’s expert contributor network publishes thoughtful, solutions-oriented stories written by innovative tech professionals. It is the tech industry’s definitive destination for sharing compelling, first-person accounts of problem-solving on the road to innovation.

Learn More

Great Companies Need Great People. That's Where We Come In.

Recruit With Us