What to Expect and How to Prepare  

With Analytics-Link’s Andrew Jones 

Data science is the hottest industry in the world right now and even though the supply of qualified data scientists remains outstripped by demand, it’s still a highly competitive field. So, if you want in on what Glassdoor has named the #1 job in America four out of the past five years, you may find the biggest hurdle to your success is getting in the door in the first place.  

If you’re looking to transition into a career in data science, there’s a lot of preparation you can do in advance to make the job hunt easier for you. Obviously, getting a Bachelor’s or Master’s degree in data science will help prepare you for the role and prove your qualifications. But while learning the technical skills data scientists need to do their job effectively is clearly paramount, it’s not the end of the story. You still need to ace the interview, and doing so may not revolve as exclusively around technical skills as you may believe.  

That’s why Discover Data Science talked to Andrew Jones, who’s been working in data science and analytics for 14 years alongside companies like Amazon and Sony, where he prototyped machine-learning features for the PlayStation 5 that Sony eventually patented.  

In March of 2020, Andrew embarked on a completely new journey, creating his own online data science learning program called Data Science Infinity. “Students learn the right content, the curriculum is based on conversations with hundreds of leaders, hiring managers, and recruiters, so they know they are focusing their time learning skills that hiring managers actually need and want,” Andrew says. “The students learn in the right way, with a strong focus on intuition and understanding. They learn and apply concepts hands-on, using real world project-based applications.” The program (and thus the students themselves) is designed to evolve over time, and students receive unlimited access to not just current but all future content. 

Discover Data Science felt this experience made Andrew the perfect person to talk to about what to expect and how to succeed when interviewing for a data science career, including what questions to ask, what questions to expect, and how to handle any testing you might be required to do. Here’s what he had to say! 

Why is data science a good career move?  

Data science is most definitely a good career move. It’s exciting, it drives a lot of value for companies (in virtually all industries), plus it’s lucrative and future-proof. After all, data isn’t going anywhere and so neither is data science.   

What kind of background, education or experience is needed for a data science role?  

The reality of this is quite different from what you might think initially—or what you might have been told. For those looking to transition into the field it can seem very daunting at first. Job descriptions are littered with requests for PhDs, or at least a Master’s Degree in data science or computer science, not to mention an ever-growing list of skills and tools that just seem out of reach.   

But, let me tell you right now — no matter your background or your qualifications, a move into data science is 100% doable, and you are capable of achieving great things in the field. I came into the field via a connection on my cricket team, which is about as left field as it gets, but I’ve worked my way up from there to work for some of the top tech companies in the world. 

Job descriptions don’t always reflect the reality of what companies need and want – so take them with a pinch of salt. Like learning any else in life, learning Data science takes enormous amounts of patience, commitment, dedication, and curiosity.  Despite what many boot camps or courses will promise you, in reality there is no “learn data science in 6 weeks” or “become a data scientist in 3 months.” Data science is an endless pursuit. No one, not even those at the top of their respective fields can keep up with the pace of the industry — but this is what makes it so exciting! 

If you are considering a move into data science, I will offer this advice: Do it for the right reasons. Do it because you love to build things that matter – complex or simple. Do it because investigating patterns in a dataset titillates you more than it should. Do it because learning about amazing concepts & algorithms fascinates you, excites you, and drives you to learn even more tomorrow because it just didn’t quite make sense today. Do it if you don’t mind being challenged and if you can take a rejection or two along the way. These will make you stronger. Do it if you’re happy to not know everything because no one does or can. There is no destination, just a journey. 

Just don’t do it for the money, or for the “data scientist” title — you won’t get far with those aspirations alone. 

What coding languages should you focus on learning? What about ML algorithms and Deep Learning?  

When you’re starting out, my advice on programming languages would be to learn SQL and Python. Why SQL?  Well, even though it’s not quite as “sexy” as languages like R, or Python – the reality is that it’s still the foundation for solving so many business problems. 

While we see a lot online about the incredible Deep Learning solutions at the cutting edge of the field, in reality, most companies still have most of their data stored in relational databases, so SQL is a must for extracting, manipulating and analyzing this data to get to quick and actionable business insights. 

To back this up with some data, I’ve spoken to hundreds of data science leaders, hiring managers, recruiters, and key stakeholders to ensure that the content for Data Science Infinity represents the demands of the field as closely as possible. 

When I asked data science leaders if their data science teams used SQL, 97% said yes. This is resounding. If you want to work in data science or analytics it is a must have skill. When I asked that same question about Python, it came out that 87% of Data science teams were using it, compared to 55% that were using R. Both Python and R are fantastic, but I suggest Python purely because it’s being used more frequently. 

In terms of machine learning, yes, it’s a pretty big part of data science, so learning how to apply the commonly used algorithms is worthwhile, but my main advice to those building up their skills is: Don’t think you need to learn them all! In the real world, you can solve the vast majority of business problems with a small subset of algorithms, so focus on getting a deep understanding of a few, rather than a high-level understanding of them all! 

As for deep learning? Yes, definitely! Deep Learning is amazing but it can also be something of a temptress! Don’t jump straight into deep learning. Don’t skip the foundational skills first. I have seen this time and time again, and it doesn’t translate well into the real-world job market! Get the core skills first and then take on Deep Learning! 

Moving away from the tech skills, what else should you have in your toolbox in terms of soft skills?  

I’m so glad this question has been asked! It is so important to emphasize that data science is not all about technical skills. In fact, the best data scientists I have worked with in my career are not the “smartest” ones, in terms of technical ability. Sure, they know their stuff in terms of coding, statistics, and other key data concepts, but what differentiates them is that they understand what the business problem is, or what the business is trying to achieve. They use data, and their unique skill set, in clever and often simple ways to solve these problems or to add tangible value to the team, business, or end-user. 

Good communication skills are vital. A good data scientist knows a lot of technical concepts, but a great data scientist can simplify these down in a way that gets everyone in the business onboard. 

As data scientists we’re here to solve problems, not introduce new ones. We’re here to enhance, and accelerate business decision making, not get in the way of it! No one is going to pay you just to be good at coding, or just to be good at math, or just to know a lot of machine learning algorithms – but they will pay you, and they’ll pay you well, to add tangible value to their business or to the end-user. 

What does a data science interview process look like, and what’s a good way to prepare?  

First, your Resume or CV will be screened (and perhaps your portfolio of projects), and if you have the required skills and experience, you will have an introductory chat with HR or a Recruiter. From there if things get more serious, you’re likely to have a phone screen with the hiring manager and this can also often include a technical or coding test. In some cases, you will be given a take-home assignment to complete. If you ace that, then you’re heading to an on-site interview where you’ll most likely meet a senior member of the team, the hiring manager, and a selection of people who are stakeholders to the role in question. 

This can all seem very daunting, especially when you stop and think how competitive this field is. Hundreds will often apply for each data science role so it seems like the odds are stacked against you from moment one, but there are some simple ways to move ahead of a lot of the competition. This is a huge part of what I help my students within the Data Science Infinity program. 

You’re essentially looking to showcase your skills at all times. This is on your resume, within the projects in your portfolio, and when in the actual interview chair. Don’t think of these as that different from each other. Think about them as all being part of one representation of you and the work you’ve done, just either at a really summarized level as on your Resume, or in much more detail in your projects, or in interviews. 

For each of your projects, when preparing for the interview process, re-work it into the STAR format.  This is an extremely simple, but hugely impactful way to deliver your work in a logical, free-flowing narrative whether it’s a project on GitHub, or when the words are flowing out of your mouth in the interview. 

If you’ve not heard of the STAR format, you’ve essentially got… 

S = Situation: The context around the business problem and why it needed to be solved – this really pulls the interviewer or reader into your narrative.  It gives a foundation for you to go into all the details without losing them later… 

T = Task: What was your specific role in the project – super simple, but important. 

A = Action: The specific actions you took from inception to conclusion.  This is where most of your answer or work will lie.  For each project, try to refine this down to the most succinct & compelling narrative you can, but keep supplementary context up your sleeve, for example the justification for why you chose solution C over solutions A and B.  A good interviewer will ask this – and it can really show you’re an expert. 

R = Results:  Super important – but often missed or under-emphasized.  You really want to show the impact that your work or your project had, this exhibits to a reader or interviewer that you can add value to their team or to their business.  Use tangible figures to add even more credibility to your work. 

Something that can set you apart from other candidates is, for each project, asking yourself: “If I was to start this project again now what would I do differently?” 

This sort of thinking can be so much more impactful than you might think – it shows you have an awareness of business impact, it shows an understanding of the nuance of what you do from a technical point of view AND it shows a growth mindset, that you’re always looking to improve.  And, trust me, that is a lethal combination! 

When you’ve documented each project in this way, simplify that down for your Resume.  Be sure to include one sentence on what the business problem was, one sentence on what you did, and one sentence on what the results were. 

Structure your portfolio projects in this way – again, it gives a really nice flow that makes it easier for a recruiter or hiring manager to see you have a clear and powerful thought process. 

And when preparing for interviews, practice speaking through in the STAR order.  Like I say, it’s so simple, but can really help you stand out from other candidates! 

How would you approach a coding test or a take-home assignment?  

My research for Data Science Infinity tells me that around 61% of hiring managers do get candidates to undertake a coding test of some sort, 29% do sometimes, and around 10% do not. Now I would look at those percentages and say, yep, this is definitely something I should be prepared for! 

Coding tests will normally take place sometime after you’ve spoken to the recruiter or to HR so, my advice is this… Make sure you ask everything about the coding test on that call. 

Ask what it involves. Ask what coding language it will be in, ask exactly how it will be conducted, will it be done online, will it be done on an internal company portal, or will it be undertaken using a platform such as HackerRank or Leetcode? 

Get as much information as you can!  If you can be more prepared than other candidates because you were sensible enough to ask questions then straight away you’re putting yourself in an advantageous position. 

Then, practice.  There is nothing better for this type of test than being in the right mindset, and this comes from consistent practice. 

There are lots of online resources for this, but make sure you’re focusing on the right thing. Don’t get super stressed trying to answer all sorts of insanely complex software engineering type problems when you aren’t going to need that. 

Regarding take-home assignments — not all companies make you do these. From my research for Data Science Infinity, we’re looking at around 55% of hiring managers utilizing take-home assignments for candidates and 45% who do not. 

Now, even though 55% might be lower than you think, if you’re applying for ten roles then you’ll be expected to do this for about five or six of them, so it’s good to know how to approach these in a way that will impress and put you ahead of the efforts from other candidates! 

Here are some high-level pieces of advice to keep in mind! 

• Always make sure you understand the precise problem being asked of you. Clarify your questions early! 
• Think hard about why they are getting you to undertake this particular task because there will be some motivation behind it. What area of the assignment is worth the most time investment when thinking about the role in question? 
• Make assumptions about the data or their company where necessary – but state them as assumptions. 

What should you be showcasing in a portfolio project? 

Portfolios can be an effective way to highlight your skills when you are early in your Data science or Analytics career. 

But I want to quickly bust a myth about these projects, and that is that Data science portfolio projects need to use huge volumes of data and/or use extremely complex Machine Learning or Deep Learning solutions! Trust me, they really do not. 

A varied portfolio of clearly communicated projects can be much more beneficial – so when building up your portfolio, look to put much more emphasis on clarity & impact rather than just on complexity. 

Just like with resumes — the reality is that hiring managers are super busy with the day-to-day tasks of their actual job, and recruiters may have more than one hundred candidates to screen. You need to make it quick and easy for them to see your problem-solving style, and your value – and “easy for them” is not screeds of code, or some super complex algorithm without any context around what it is actually solving! 

What sort of questions should you be prepared for in advance? What is the most common thing an interviewee is going to be asked? 

In the final stages of any interview process, you are going to get more “behavioral” based interview questions. 

Part of my research for Data Science Infinity was to ask hiring managers what their “favorite” interview questions were — the ones that were most useful for them in terms of getting closer to a candidate’s abilities. When looking through these, there were some common themes, so I’ll share some example questions that cover those themes here. 

1) Tell me about a project you have worked on recently and explain to me what the commercial impact of that project was. 

2) Tell me about the piece of work are you most proud from the last 6 months – and explain what you would have done differently if you could start again 

3) Explain how the success criteria were defined for a project you worked on. Why were these metrics important, and what was the end result? 

What you will notice is that these questions are always broader than the technical details of the project itself. Question one refers to the commercial impact, question two refers to your growth over time, and question three is focused on the success metrics surrounding a project. 

Long story short: Hiring managers want to see how you can add value to their team and to the business, and they want to see that you can think holistically about your projects, right from inception through to implementation and results. 

Prepare for interviews in a way that covers each project as a whole – I guarantee this will set you apart from other candidates! 

Any common mistakes interviewees make? What does an interviewer not want to hear? 

The most common way that candidates keep themselves from getting the role they want is exactly what I’ve spoken about already. Cover your project in a narrative that takes the interviewer right from “what was the business problem, and why did it need to be solved” all the way through to “this is what the outcome meant for the business or the customer.” You are being hired to solve problems and add value to the team and business.  If you can show evidence of this in your past, that will translate well to success when going for new roles. 

Something that isn’t covered often is what to ask your interviewer. There will usually be a time, most likely at the end, when the interviewer will ask “Do you have any questions for me?” You should have questions because this is actually a really key part of the interview, and it can be a good opportunity to end on a positive note. It is obviously a great time to get clarity or further information about things you’ve discussed, so make sure you ask about anything you want more detail on. If you don’t need clarity there are still ways you can use this time very effectively. 

Do not ask: “What don’t you like about working here?”   

Are there most likely things that they’re not happy about in their work? Sure. But they probably aren’t going to tell you about them in great detail. All it does is make them think about something negative, and we really don’t want any negative associations with the time they spent with you. It can also make it seem like you’re actively looking for reasons not to work there.  

Do not ask: “How did I do?” or “What are my chances?” 

Often an interviewer will want to absorb everything and speak with other interviewers before making up their mind. They don’t want to be put on the spot like this and, if you did poorly, they’re not likely to admit that to your face anyway.  

Instead, try to remember the topics the interviewer seemed passionate about during your discussion, and ask them for more details on these. If you can’t remember what they might have been excited about (or if they didn’t speak to that) then ask them that very question: “What projects are coming up in the next 6 months that you’re really excited about.” 

With this question, you’re asking about an interesting project that the very team you’re applying for will likely be working on. Not only is it a relevant question, it makes them speak about something that excites them.  Remember, your overall goal in the interview is to show them that you can add value to their team and to the business, but you’re also looking to ensure that they enjoy their time with you – and ending the interview on a positive note can be a really good way to do this! 

Learn more about Data Science Infinity and discover how you can get your own data science career jumpstarted with guidance, support, and mentorship from Andrew. You can reach out to him on LinkedIn, where he posts daily, to connect further. 

___

Andrew Jones, Founder/Director of global data science hub analytics-link.com, is a published author and mentor who has interviewed & screened hundreds of Data Science candidates throughout his career. Developer of several machine learning patents for Sony PlayStation regarding Machine Learning solutions, his near decade and a half in the industry enables him to provide unique guidance, preparing candidates for success with interviews, placement, and promotions. You can learn more on this subject by visiting his website or in his book The Essential A.I. & Data Science Handbook for Recruitment.

Return to Discover Data Science Articles