The Skills You Need to Tackle Data Science Interviews as A Fresher

“I’m a final year student enthusiastic about data science and want to start my career at XYZ. I have done a few data science certifications; what other skills are important for becoming a data scientist?”

This is a common question I repeatedly receive in my LinkedIn inbox. It was primarily directed at my previous workplace, but you would have wondered what it takes to join the team or become a data scientist in general.

I felt answering this once and for all would help all data enthusiasts, and I called my manager and asked him the same question: What skills would you look for in a fresh graduate who aims to become an entry-level data scientist?

It was a fruitful discussion between us. We discussed what we felt was missing among the candidates and what is non-negotiable and good-to-haves. So essentially, what follows is the summarized version of what we think are the most critical skills required to become an entry-level data scientist.

1. Master the Fundamental Pillars

When we learn every new trend in the AI industry, such as the latest deep learning framework, we forget to build our basics. You must understand the underlying concepts behind data science and master them first.

Most universities indeed teach them, but you might want to refresh your memory. You don’t need to have a master’s or even be in university to learn these basic concepts; they’re available for free online.

Mathematics for Data Science

All machine learning concepts rose from basic mathematical concepts, and you need to focus on having a thorough understanding, at least of the backbone of machine learning:

• Linear Algebra from Khan Academy
• Multivariable Calculus and Derivatives from Khan Academy

Statistics for Data Science

Unless you come from a direct statistics background, this is the topic we see most people skip. But when we look at our day-to-day work, topics such as hypothesis testing, type I and type II error are crucial. Please do yourself a favor and learn enough statistics before it’s too late.

You may find the Statistics with Python Specialization from the University of Michigan relevant.

Programming for Data Science

All data scientists eventually need to code. You don’t have to be a pro in multiple languages, but you need to be able to create the basic pseudo code when given a problem to solve.

Most companies use Python, PySpark, R, and SQL, so having some experience in one of them would make your lives easy. Research which of these languages are commonly used for your company of interest.

Refer to this article based on your preferred programming language for resource recommendations.

2. Learn Sufficient Machine Learning

You probably saw this coming, and once you’ve mastered the fundamentals, move on to various machine learning algorithms. Go hands-on but also understand the underlying concepts behind each algorithm and when to use them. This is the exciting part of the day-to-day work of a data scientist.

Don’t listen to us — learn it directly from Andrew Ng.

3. Show — Don’t Tell

It shouldn’t surprise you, and most of the candidates claim to have a sufficient understanding of all the concepts we’ve listed above, and we can only test a little during the interviews.

There is a better way.

The best way to demonstrate your skills is to implement an end-to-end project using all your learned concepts. You may start with your final year project or even an independent project you were passionate about.

Show us your skills, don’t only tell us.

4. Cultivate the “Can-do” Attitude

You might think: I have all the skills in the world, a portfolio of projects to brag about, and a degree from a renowned university. But if you don’t have the right attitude, none of them matter.

Data science is all about teamwork, and each one of us plays a specific role in each project. The truth is: it’s hard to work with someone who doesn’t have the right attitude. The skills, then? You can always acquire them, and we’ve got people in the team to help upskill.

Some pointers on having the right attitude:

• Data science is an evolving field, be open to learning consistently
• Some tasks are challenging by design; only people with the ‘can-do’ attitude can solve them
• Offer and accept a feedback positively. We all want to grow as a team.

In Summary

To summarize, we looked at essential skills one needs to make as a data scientist:

• Fundamentals of Mathematics, Statistics, and Programming
• Basics of Machine Learning concepts and algorithms
• Projects where you have applied the theory learned
• The “can-do” attitude

I intentionally kept it general and straightforward so as not to overwhelm you with many interview questions and whatnot. I suggest starting with this, then looking up interview experiences for your specific company online (on websites such as Glassdoor or LeetCode).

I need to leave you with a disclaimer — while this post will help you better understand the skills required and upskill, it cannot guarantee a position in any team. Whether we like it or not, many more factors, especially business requirements and vacancies, influence hiring decisions.

The skills required in the data science industry are constantly evolving but remember this: every expert started as a beginner somewhere.

Start upskilling now; after reading this post — you could be the next success story.

For more helpful insights on breaking into data science, honest experiences, and learnings, consider joining my private list of email friends.