Recruiting for Data Science Teams

I have worked on different Data Science teams for more than 5 years now and have been on both sides of the table. In this post, I will share some of my thoughts on what to look for while recruiting for a Data Science team.

I have learned a lot from my mentors and managers with whom I have conducted recruitment interviews and processes. Their thinking has influenced a lot of my own thought process and it has been eye-opening to see how each leader approaches the process differently. Additionally, every team or organization has a different style of recruiting that is in line with their philosophy and how they operate. What I have tried to outline here are some key parameters that I evaluate during any kind of recruitment process. I will delve deeper into the recruitment process itself in a separate post.

In my experience, it is always hard to find talent with the right skills and domain expertise that is required for your team. Additionally, the platforms, tools and technologies used in Data Science change so quickly that the goalpost keeps shifting.

For starters, my approach has always been to recruit smart people and then equip them to figure out the rest. I think this is also reflective of my recruiting experience in the Indian context, where there is a narrow funnel at the top, of people with advanced mathematical/ scientific degrees but a large funnel of people with graduate level engineering degrees.

During the recruitment process, I essentially evaluate candidates on the following four parameters in no particular order —

Educational Qualifications — More than what the degree or GPA states, I would like to know what the person did during the degree program. Was there an automated SCRABBLE project that could play on it’s own OR an internship that provided exposure to some technologies that are currently used in your own team. The ranking and reputation of the University is also given a high importance along with the fact that the degree was of a technical nature and involved coding and logical ability. Interestingly, I have observed that this parameter is very important when recruiting for a junior level position but matters less when you recruit for more senior positions.

Projects — I use this parameter to understand the depth of knowledge that a person has. Does she/he understand the nuances of the work they’ve done and is able to articulate why a particular choice was made? Have they taken time to evaluate other approaches and if not, are they able to think about them during the interview? It’s always interesting to observe how a candidate reacts when an alternate approach to their own project is presented. Do they understand the overall architecture of the project, where their work fits in and the end goal or business objective? I’ve often seen that candidates who understand the big picture are able to work through assigned tasks and specific problems faster and without boiling the ocean.

Communication — By this I definitely do not mean a test of language skills! This is to judge the candidate’s ability to understand and assimilate the business problem. She/he should also be able to re-state their understanding of the problem and articulate their approach and solution. I’ve seen candidates who may not be the best in expressing themselves but are excellent at grasping what we’re trying to solve and quickly work towards it. The ability to work on a team, understand your team-mates and work together is one of the most important aspects of a data scientist. I have also experienced that most of my learning and self-improvement has happened through what I have learnt from my colleagues and therefore being able to understand and communicate is paramount!

Attitude & Learning ability — I said earlier that we need to recruit smart people. Alongside we also need to ensure that they have the right attitude towards learning. Is the candidate curious and is there a desire to learn? Are there some signs that exhibit this quality? For example, participation in Kaggle competitions and other hackathons. There is always a ton of stuff to learn in any new role — domain, techniques, technologies, working styles etc. We need team members to be open to learning new things every day and not rest on their laurels. There needs to be excitement about learning and a fire-in-the-belly attitude towards achieving success. This is sometimes reflected in the candidate’s self-awareness — does the candidate know what she/he is really good at? Are there some clear focus areas they want to work on? I’ve seen that candidates at the junior level are not as self-aware as more experienced folks, nevertheless these questions when asked force them to think and assess themselves.

Over the years, I have realized that I enjoy being part of recruiting discussions because I get to meet some interesting and talented people who could hopefully become team-mates. I will also be writing a follow-up post with details about the recruitment process. I hope that this post was helpful and do let me know what you think.