Data science interviews: the good, the bad and the ugly

The good

An interview is often a lengthy and tiring process (sometimes even nerve-wrecking, but we’ll get to that later ?!), so having an interview that doesn’t feel like a total waste of time is nice! But, according to our data science team, what actually makes an interview good (or even great!) are several aspects:

An interview is a two-way street

Although the interview is mostly focused on assessing if someone is a good fit for the company they are interviewing for, everyone attending the interview should remember that the interview works both ways.

1. The interviewer will assess the skills and/or cultural fit of the interviewee
2.  The interviewee will have a chance to understand whether they would like to work with the person interviewing or within the company. 

Being in the position of the interviewee should not mean you are a passive participant, and only there to answer questions. Instead, come prepared with a list of things you want to find out about the role, the team and the company. Take notes throughout the interview and ask questions, where needed, to fill in any gaps. Not only does this show your genuine interest in the position, but it will also help when it comes to making a decision between multiple roles.

A great start is promising

An interview is the kind of experience that usually makes most people nervous, especially the interviewee, but more often than not, the interviewer is nervous as well. Some of the best interviews we have experienced have started with some small talk, perhaps a joke; anything that doesn’t jump straight into the assessment! Finding common ground and building some rapport can help both parties relax into what should be a pleasant conversation. 

Being outside of your comfort zone may be better

This may seem to contradict the previous advice but stay with us for a moment…

Think about the projects that you are most proud of; these are probably those that challenged you and where you learned new (and maybe painfully complex) things. The struggle, the teeth-grinding, the late nights – but, in the end, it feels like a great achievement when you see it done!

Interviewers sometimes like to ask difficult, possibly impossible questions. The goal here isn’t to trip you up or make you struggle, they just want to see how you react and understand your thought process. Managing difficult situations is part of being a data scientist, and knowing how to handle situations when everything is new to you is all part of the job. One of the key skills you’ll need to succeed in this profession is problem solving. So, here’s our hot take on how to handle these sort of questions:

• Take a moment to think and ask clarifying questions (just to make sure you don’t miss anything obvious) 
• Then, provide the best answer you can think of – and own it!  Make sure you vocalize your thought process so that the interviewer understands how you approach uncomfortable situations/problems. This is more important than getting the right answers!

The bad (or weird…)

We’ve all had our fair share of weird social situations – not only in an interview environment! 

A weird start is not a show stopper

Imagine you go in, take your coat off and then you struggle to hang the coat for some good two weird minutes, while everyone in the room looks at you asking what are they doing?! Now, you manage to hang the coat, you feel the relief, take a seat and then knock over the glass of water in front of you. Imagine the feeling – the sheer embarrassment – but, trust us, all is not lost! Remember a great start is promising (the good!) so here is your chance to turn things around and show your sense of humor, make a joke and move on!

See it as an opportunity to show them how you react in uncomfortable situations. That’s a skill too! This is something that happened to one of our data scientists and it didn’t stop them getting hired! (Disclaimer: We do not recommend doing this intentionally as an icebreaker…) 

Being genuine is better than being great

It’s safe to assume that all of us try to make our CV as appealing as possible, however, everything that’s on the CV becomes subject to questions – so be prepared to chat about everything and anything you list. Before you add something on your CV ask yourself; is this something I can (and want) to chat about for a good two or three minutes? If the answer is no, then maybe it’s better you keep it out, or else you might end up talking about some project that you did a while ago and that wasn’t particularly interesting. This means you’ll miss your chance to make yourself shine chatting through a project you are passionate about!

Play to your strengths

During the interview, you should constantly try to show your best self, but sometimes you might get a question that you don’t know how to answer! Of course, it could be that even if you had more time to think about it, you would still not be able to answer it, but imagine it’s the kind of question you know you’d be able to tackle if you weren’t pressed for time! 

An example of such a question could be a short coding exercise in a programming language you are less familiar with. Your best bet is to be honest. Say that the syntax of that programming language is not familiar to you and ask if you can solve the problem in a language you use daily, or if it’s an option to do it as a take-home task after the interview.

This way, you are still on the path of showing your best self; you prove you don’t hide from challenges and that you’d like to demonstrate your skills. The worst that can happen is they say no (which wasn’t the case for the analyst who volunteered this story!)

Be wary if you’re tested on something irrelevant

Tests and technical questions are a standard part of the data science interview process and serve the important role of assessing the capability of a candidate. But, sometimes, we’ve found ourselves pondering the relevance of a particular test to the role we’re interviewing for. 

At best, irrelevant tests are a redundant part of the process which should be replaced with something more role-specific (hiring managers, take note.) At worst, irrelevant assessments could be major red flags and be representative of the true day-to-day nature of the role. Our advice would be to always try and complete these to the best of your ability, but don’t be afraid to query the relevance of the test to the role that you applied for.

The ugly

Bad interviews are one thing, but sometimes it’s the process that is broken. Going through a good or a weird interview is easier than going through a truly awful one. Unfortunately, not all experiences are the same, so we think it’s important to discuss the ugly parts of it as well. 

An awful experience is not the end of the world, and it often doesn’t say anything about our capacity for solving difficult questions or integrating within a team; it’s more about incompatibility between the two sides. However, when you realize that the negative outcome was mostly because of how the interview was planned and led, it definitely feels like a setback and, potentially, a waste of time. In this section we’ll cover the ugly experiences that we’ve had as a result of poor processes. For much of this, there may be little that you can do as an interviewee, but we think it’s important to be aware so that at least it won’t throw you off your stride too much or knock your confidence when job seeking.

Long, drawn out interview processes

Sure, it is important to find the right person for the role – but did they really need five interview stages plus technical tests for a junior role? Probably not. Sometimes a process may happen over the space of weeks, maybe even months, and you’ll have a never ending list of people to meet. When this happens, it’s important to remember not to put all of your eggs in one basket. Don’t stop your job search hanging on for that call back, – keep applying and keep interviewing elsewhere in the meantime.

Three-in-one jobs

It’s important to understand what the main focus of a role is when applying. Sometimes, companies are looking for a data scientist that can do the job of a data engineer and a data analyst as well. More often than not, this means that you’ll spend more time figuring out ingestion issues and/or building data analysis reports, rather than on solving those data science problems you are hoping to tackle. It’s essential that you understand what a day in the life of a data scientist looks like for the company you are joining, and whether that’s how you want to spend your time.

Record-yourself interviews

Our colleagues in the marketing team don’t believe this is a thing, but it’s happened to a few of us in data science. Collectively, we all hate interviews where you have to record a response to set questions. Completely impersonal, with formulaic questions and with no opportunity for feedback, it is nearly impossible to demonstrate the amazing asset you could be to a business when recording responses to set questions.

Generic rejection

These are a common gripe with interviewing, regardless of industry. These emails offer you no feedback on what you could do to improve for the next interview. Ah well, at least you got an answer and you didn’t just get ghosted! Although it’s disheartening to keep on waiting and not really knowing if you should be waiting any more or not, this actually says more about that company than anything else. Chin up, onto the next one! At least you know you can cross this one off your list, which brings us onto our final ugly experience…

Red flags shouldn’t be ignored

It sometimes happens that we’re so sure a certain job is for us that we deliberately ignore the red flags that come along the way. The worst thing that can come out of an interview process is for you to start a new role only to realize that it’s completely wrong for you for whatever reason, and that all the signs were there during the interview process. For instance, imagine receiving a take-home task, spending a lot of time and effort on it, and then barely getting feedback on it. This may mean that you could be joining a team that doesn’t value other people’s work. Be mindful that, although it’s a relief that you passed, it may not be a good sign. 

To conclude, we’d say that finding the perfect data scientist job, and one that you’ll absolutely love, is not easy! However, the interviews are a good hint on what to expect if you were to join a company. It’s tremendously helpful to know, from the beginning, what you’re looking for (and what you want to avoid) in a company and in a job. 

Good luck with every endeavor you may pursue, we hope to see you at an Ensemble event soon to swap more interview stories!