Assessing candidates for the right skills for data science positions can be a daunting task. While recruiters may think the outcome of this assessment depends on the depth of technical knowledge candidates may have, there are other things that interviewers should consider which is essential for business stakeholders.
In this article, we will take a few facets that data science recruiters may be missing in the hiring process.
Evaluate Your Data Science Hiring Process
According to social media posts from data scientists, the large majority of companies that hire data scientists don’t have much clue on how to properly assess data science skills. Those companies may not have the necessary know-how on how to assess the capabilities of data scientists on various business problems. So, it is very important that companies consult with hiring firms or specialised organisations that are apt at assessing the skills. Interviews can also be taken by other data scientists within the company as candidates would likely work with them.
Does Your Hiring Manager Know What To Ask?
The problem is that many first-round interviews are done by staff members, particularly HR professionals who may not have many ideas on how to properly assess data science candidates. Even IT professionals and C-Level tech executives like CIOs/CTOs may not have much experience in core data science applications, and the questions could focus on tools and techniques, instead of understanding the data-based aptitude of candidates to solve business problems. To get around this issue, senior data science team members can be asked to take part in the interviewing process.
It’s also important for such businesses to take a post-interview survey to know whether the questions were fair in assessing the right aptitude or whether the questions were too generic. Similarly, companies can hire specialist recruiters who have a deep understanding of the data science aptitude, and use their expertise to conduct interviews. In many cases, the in-house HR staff is more likely to be incompetent when it comes to data science-related hiring.
Ask Whether A Candidate Understands The Business Problem
Candidates should be interviewed on their ability to study specific business data points that have a huge impact on decision making in the dynamically changing world. At the end of the day, prospective data scientists are expected to generate money for the company, and without the ability to look for the right data, it is going to be a failed attempt at the data science hiring process. Instead of the theoretical knowledge on data science techniques, one of the key aspects of the interview process should be focused on problem-solving that starts with how to get the right data, not how to build a model.
Assess Not Just On Techniques But Thought Process That Can Help Your Business
There are thousands of techniques in data science, and every week there is some new innovation in AI/ML. So the focus should be on the fundamental concept of data science. During an interview, instead of asking candidates about data science theoretical concepts, and generic questions on tools, the candidate should be assessed on practical responses to real-life business problems and specific projects. But, at the end of the day, theoretical concepts are at the tip of your hands, and just one Google search away.
Also, tools like AutoML can apply all sorts of algorithms on data automatically. If that’s the case, why do you need data scientists anymore if it is just about using tools? On the contrary, machine learning is just one part, and AutoML doesn’t cover all possibilities, and simply applying techniques is easy, but having the aptitude to find such people is the challenging part.
Understand A Candidate’s Logic Behind Using An Algorithm Or Technique
Experts say that if they want to hire people who could simply use ML tools or Python, it can be fairly easy. But, that’s not what they should be striving for during interviews. In fact, they should ask questions to know why a candidate would want to apply a certain algorithm on a certain dataset, instead of just knowing whether someone knows a tool, platform or programming language. The thought process in problem-solving is very important for data scientists. A good data scientist can come from any background, and the best data scientists may not be in the place you are looking for.