Top Qualities Hiring Managers Look For In Data Scientist Candidates
Top Qualities Hiring Managers Look For In Data Scientist Candidates
Some of these are arguably more important than writing efficient code
Some of these are arguably more important than writing efficient code
As 2021 came to an end, I reflected on my data science journey in the past year and realized that one of my proudest achievements was helping to hire two of the best data scientists I have ever worked with for the team. Their great work undoubtedly contributed to my promotion to a data science manager.
In the new year I will be the official hiring manager for my team and like one of my favorite managers told me, “hiring right will contribute monumentally to your success as a manager”; this motivated me to reflect on my past success in the hiring process as an interviewer and to codify some of the learnings from it. Hopefully this will serve as a memo to myself and other data science managers as to what attributes we should watch out for in the hiring process and also serve as a guidance for aspiring data scientists as to what qualities/skills are valued by managers and will make you a better data scientist.
Ability/willingness to learn
This is arguably THE most important skill/attribute I look for in candidates; because let’s face it, data science is a fast-developing field in which the only way to keep up is to learn continuously.
There are constantly new tools, algorithms and approaches being introduced; so it’s extremely frustrating to have people on the team refuse to adapt to new things because they are “used to” or “familiar with” something else. I frequently see this fear of needing to learn new things cloud people’s judgement — for example, the fear of needing to learn Looker might lead people to strongly push for using Tableau. Don’t get me wrong, I’m not saying Looker is better than Tableau, nor am I saying that the team can’t push back on decisions made about the tech stack, tools and other things; but I can’t accept the basis of push back being the inability and/or unwillingness to learn new things.
How to test this in interviews:
You can test candidates’ willingness to learn and ability to adapt by asking them about a time when they lacked a skill necessary for their role, and how they dealt with that situation. Similarly, test their drive by asking about the last time when they proactively learned new skills related to their field either on the job or in their free time.
How to build this skill:
The ability to learn is like a muscle — you have to keep on training it to keep it active. The best way to do that is to keep yourself up to date with trends in the field by either casually reading relevant articles on the topic (medium is a great platform for this), or periodically intentionally up-skilling yourself by taking online classes (e.g. via Udemy or Coursera).
Speak the code as well as the business
The ability to code is crucial to good data scientists, that’s why almost every data science role has a technical round. But what’s equally important but sometimes overlooked is the ability to understand the business. Without the business acumen, data scientists will always be the passive implementer of tasks instead of the active thought partners that they should be. Moreover, only when you truly understand the asks and how they fit into the larger business, you are able to problem solve in creative ways without counting on others to prescribe a solution.
How to test this in interviews:
On top of the technical challenge, construct a business case which candidates have to work through. The business case should closely align with the job description. If the role will be conducting a lot of metrics analyses, then the case could be a metrics decomposition type of question; if the role will be mainly building models, then the case could be a realistic business situation that candidates can brainstorm modeling solutions for.
How to build this skill:
The best way to build this skill is stepping out of your immediate scope and trying to learn as much as possible about others’ work in the company. Talking to different people in different roles, learning about their job and thinking about how their roles fit into the company is the best way to build up business acumen in my opinion.
Additionally, familiarize yourself with your company’s strategy and practice asking yourself how each of your pieces of work ladders up to company priorities. And lastly, stay up to date on industry news and developments so that you have business context beyond your company’s immediate roadmap.
Stakeholder management
Companies nowadays are trying to enable data-driven decisions throughout the business by having data scientists either embedded in or working closely with various different teams. So managing stakeholders has become a crucial part of data scientists’ day-to-day job. What aspect of stakeholder-management is important might vary from role to role, but there are several that I have seen being universally crucial based on my experience.
For most data science/analytics asks, it’s important for data scientists to work with stakeholders and translate business needs to into analytics needs; so the ability to listen to the need of stakeholders and to dissect a business problem is fundamental to a good data scientist.
Stakeholder management is not about saying “yes” to everything; on the contrary, it’s important to know when to say “no” and how to communicate “why not”. The ability to manage stakeholders’ expectations and explain the limits and caveats of analytics approaches is key to being a good data scientist.
How to test this in interviews:
The best way to test this is by having cross-functional stakeholders on the interview panel and assess the candidates’ ability to interact with them, allowing key stakeholders to see wether the candidate is a good fit and a person they want to work with.
Traditional behavioral interview questions can provide additional signal (e.g. asking candidates about a time where they pushed back on a stakeholder’s request, and following up to understand why and how they did it).
How to build this skill:
The best way to build up this skill is to seek cross-functional opportunities and work with as many people as possible, and try to completely “own” the relationship without having your manager act as an intermediary. In the process of working with different people, really hear their needs and brainstorm with them the best analytics approach to a problem and reflect on each project afterwards. This last step is crucial; make sure that you solicit honest feedback from your stakeholders so that you can learn to become “great to work with”.
This is not something you can read off of a book in a week, it will take practice and time; so be patient with building up this ability.
The ability to identify inefficiencies and propose solutions
I have experienced my fair share of managers/leaders who like to be surrounded by “yes people” over the years; we all know what happened to the emperor who did that —ended up naked in public. And that’s the kids-friendly version where only the lightest consequence was mentioned. I want to surround myself with people who are not afraid of questioning the status quo and are not afraid of proposing changes and implementing them.
The infamous “imposter syndrome” is common among less-tenured data scientists; but it’s important to realize that no matter how junior you are, you could have something to contribute. So don’t be afraid to speak up when you see something inefficient, and it would be even more appreciated if you brainstormed a solution for it.
How to test this in interviews:
This one is hard to “test” in an one-hour conversation, so the best way to get a sense of this is by directly asking a candidate when the last time was that they identified an inefficiency and proposed a change. But make sure you dig deeper in the answer to understand the exact situation in order to avoid a superficial answer.
How to build this skill:
Never take anything as a given. Question things that seem inefficient to you, regardless of your tenure. In order to be able to propose solutions, it’s helpful to utilize some of the suggestions mentioned before — keep yourself updated to the newest trend in the field and also talk to your stakeholders to understand their pain-points and the best solution fitting for the situation.
Key takeaways:
Non-technical skills that will signal to hiring managers that you are a good data scientist
The ability to learn new skills and adapt to change
Having business acumen
Being able to manage stakeholders
The ability to identify inefficient processes and propose solutions
Want to read more about data science and business? I might have some suggestions for you:
These Mistakes Could Easily Ruin Your Data Science Interviews
Don’t let your onsite performance undermine all of your interview preparationtowardsdatascience.com
5 Mistakes I Wish I Had Avoided in My Data Science Career
I learned these lessons the hard way so you don’t have totowardsdatascience.com
Why I Won’t Participate in the ‘Great Resignation’ For My Side Hustle
and why you should reconsider, too, if you are thinking about itmedium.com