In this article, I’ll give you a structured approach to getting a data science job.
In fact, I’ll be sharing all the techniques that have helped me get offers from startups and management consulting firms, along with examples of my own resume and project portfolio. Additionally, I’ll talk about what I look for when screening CVs and running interviews.
So if you want to get a job in data science, you’ll love the actionable steps in this guide.
Let’s get started.
Be eligible for a data science job
Before going into the details of job hunting, let’s get this out of the way: no amount of tricks will get you a job if you don’t have the required skills. So the first step is to learn the fundamentals: coding in a data-friendly language (ideally Python or R), some machine learning, and SQL. Those are the basics for an entry position in data science. And they can be learned for free on the internet.
If your market is very hot or you’re looking for internships, you may get hired with a technical degree (CS, math, physics, engineering) and no specialized data science knowledge.
Some work experience may also help to make you a more attractive candidate. Adjacent positions like data analyst and data engineer can help you move to a data science position.
Additionally, some domain knowledge of the industry of the companies you’re applying to will be a great asset to your job search.
You have to present your story in the best light possible
And this is true through the whole hiring process, from your resume to interviews.
When you are looking for a job, you are both the product and the salesperson. No one else but you will highlight your qualities. There are many ways to explain who you are and what kind of work you’ve done in the past. You should choose the most persuasive way in every situation. To do this focus on two main principles:
- Make your story as interesting as possible by being specific enough
- Adapt your story to your audience, highlighting what is more relevant to them
For example, when Mary is asked what she does for a living in an interview she can say “I do customer segmentation” and that can be as true as it is boring and unspecific. But she can do better. She could say “I use algorithms to segment users according to their past purchases”. That sounds more interesting.
Moreover, if Mary is on an interview with a software engineer, she can specify that she uses a mix of SQL and Python code for her analysis. If she is talking to the marketing manager, she can explain how her segmentations helped increase the email open rate by 12%.
Additionally, she should try to use her audience’s own words. Her official job title states “Business Analyst”, but she’s using SQL and Python to do her job. If she’s applying for a “Data Analyst” position, she could say her current job is a “Data Analyst” position too.
These ideas apply to interviews as well as the wording of your resume and any other document you present.
Improve the steps of the funnel where you’re weak
A job search is like a sales funnel. You find some job postings and apply to them. Some of those applications will get you interviews. And some of the interviews will result in job offers.
By thinking of the process as a funnel, you can isolate its parts and try to optimize them separately. For example, imagine John has sent lots of applications and isn’t getting any interviews. In that case, before sending more applications John should make sure his CV is well-formatted and that he is a good fit for the positions he’s applying to.
The main parts of the job search funnel are:
Applications, increasing the funnel input
The first step to getting a job is finding job postings. The main ways to do so are:
- Asking your network, which may even let you skip directly to the interview phase
- Online job searches, I’d suggest searching about once per week (LinkedIn has by far been the best for me)
- Improving your LinkedIn profile and setting it as open to work
- Local job banks
- Company jobs page if you’re interested in specific companies
- Cold emails to people in your industry
Once you’ve got some job pots you have to decide which ones to apply for. My rule of thumb is to apply to whatever job you feel confident you can do, regardless of requirements. Very often companies will post job offers where it’s almost impossible to find a person with 100% of the requirements. If you don’t fulfill some of them but feel like you can pick them up easily on the job, just apply.
Maybe you have confidence issues and feel you may not be worthy of the job. In this case try to see what kind of jobs did your classmates get, and shoot for something on that level. If people that studied with you did it, you can do it too!
Unless you’ve applied to at least 20 positions, your best bet is to keep sending more applications. Think of it this way: if 20 people apply for a job your base probability of getting an offer is 5%. Lately, on LinkedIn, I’ve seen many postings with as many as 100 applicants.
Improving your LinkedIn profile
On my last job search, I got contacted by many recruiters that found me through LinkedIn and brought relevant offers. This will probably happen more and more as you advance through your career.
But to get noticed you have to work on your profile. Here’s what I did:
- Create a complete profile with all your relevant work and learning experiences
- Follow LinkedIn’s advice to improve your profile
- Write an “about” section that sounds professional
- Get your friends to endorse you on the necessary skills for your job search
- Take LinkedIn’s skill certificates to make your profile stand out
- Accept random connections, sometimes they have offers for you
- Set your profile as “Open to work”
Successful applications and getting noticed
Once you’ve selected some job posts, you need to send the best application you can. Your objective when applying is to convince HR that you’re a good fit for the job.
Before going into which documents to send, let’s talk about referrals. This is the step in the process with the highest potential return on your effort. If you have a connection working at the company, talk to them and ask for a referral. This will bring extra attention to your application and possibly let you skip some of the required steps. Just do it if it’s available.
Whether or not you can get a referral, applying for a position involves sending some documents that show who you are. Any document you include, make it a pdf. Word and other editable formats may visualize differently on different computers.
How to write your CV
The most common document is your resume or CV (Curriculum Vitae). The objective of your CV is to communicate who you are to the recruiter. Here are a few guidelines about how to write your curriculum:
- Use a nice template, for example
- Only include a picture if you look good on it, in the picture dress for the job you want
- One page, include only relevant information
- Make everything easy to understand
- Don’t get too technical (this will probably be filtered by someone with no technical knowledge)
Here’s the resume I used the last time I was looking for a job (2021).
In addition to your CV, you may want to include a cover letter or a project portfolio.
The cover letter should be about why this job is a good match for you and why you are a good match for the job. It could increase your chances of getting a positive response, especially in more formal recruiting processes such as management consulting. You can send it as an email or as an attached pdf (max 1 page). You can structure your cover letter as follows:
- Introduction: why you’re writing this document
- Why the company is a good fit for you, for example: it’s a market leader, very innovative, you personally use their products, …
- Why the position is a good fit for you
- Why you are a good fit for the position and how your previous work experience and education has prepared you for this job (try to address all the points in the job description)
Here’s an example of a cover letter I used some time ago to get into management consulting.
Project portfolio tips
Another document you can send is a project portfolio. This is a document explaining some of the projects you’ve worked on. If you have done some projects that you’re proud of, this can make them shine. In fact, the last time I applied for a job I impressed some of the interviewers with my project portfolio.
If you do so, keep in mind the following points:
- Don’t make it too long: 2/3 projects, 10/15 slides max
- For every project, explain the technology used, the process, and the results
- If possible, showcase projects relevant to your potential employer (same industry, technologies, or modeling problems)
- Don’t assume the reader has previous knowledge of your projects, give all the necessary context
- Don’t share any confidential information
If you send a GitHub link, make sure to keep your profile clean and organized. Also, create a clear readme file on your projects with a summary. Otherwise, reviewers may not know where to start and will just skip it.
What I look for when screening applications
At my past job, I screened applications of potential candidates for data science consulting jobs. This is what I looked for in order of relevance:
- Evidence of proactiveness and problem-solving in previous work experience and side projects
- Fundamental data science skills (ML, R/Python, SQL). I personally don’t care much whether they’ve taken a master’s degree or some MOOCs
- Numerical and coding skills (technical degree, side projects, …)
What if you don’t get answers to your job applications?
Looking for a job can be tough, especially when companies and recruiters ignore you. Don’t despair. If you find yourself in this spot, here’s what could be happening:
- You haven’t applied to enough jobs or have been unlucky so far, send more applications
- Your resume is poorly formatted or difficult to understand, work on it
- You have the skills but not the credentials, try to explain better why you are a good candidate and maybe give recruiters some proof of your skills (for example: portfolio, LinkedIn certificates)
- You don’t have the necessary skills for the positions you’re applying to, in which case you should level up your knowledge or apply to other more suitable jobs
Proving yourself: Tests and assignments
If after sending your application letter the company likes you, they will contact you. At this point, some companies will assess your skills and commitment with an assignment or a test.
Tests are like exams. You will have limited time to answer a series of theoretical questions or practical exercises. Ask about it and try to prepare in advance. Doing similar tests from other companies or preparation websites will help. Also, try to schedule it for a time when you’re rested.
Here are some resources to prepare for a Data Science test:
Assignments are small projects that may range from about 2 to 12h. Some assignments are downright abusive. If you aren’t too interested in the job, now is the time to get out of the process.
Other assignments are interesting and fun challenges. You can take them as a chance to see what the job will be like and also test your skills.
Whatever the type of assignment, always think that it’s a relatively small investment compared to the time you’ll spend at the job if you end up with get an offer.
What to do if you are not passing the tests
Don’t worry if you get turned down at one test or assignment, flukes can happen. And sometimes recruiters don’t know what they’re doing.
However, if you fail at tests repeatedly, that means you should review the theory. Try to remember which questions you missed and study those topics and adjacent ones. Also, try to practice doing tests, as that always helps.
If you’re having trouble with assignments, then it’s a matter of practice. Try to do projects on your own and explain them to friends. Reviewing projects by others can also help.
Interviews come in many shapes and forms but tend to follow a common pattern. Most of them will consist of 4 main parts: introduction, HR-type questions, technical questions, and your questions. Preparing for each of the parts will improve your odds of getting a job offer.
Before an interview, you should review the job description and make sure you understand all concepts mentioned in it. This will automatically make you a better candidate. Bonus points if you think for some time about their business and how data science can improve their bottom line.
In the first part, the interviewer will present herself and also give some information about the company and the role. Then she will ask you to introduce yourself. You should prepare your introduction and practice it in front of a friend to project a better image of yourself.
HR-type questions are usually about your motivations, your character, and your soft skills. Some typical HR questions are:
- Why did you decide to apply to this role?
- Tell us about your strengths and weaknesses
- What do your colleagues think of you?
- Can you describe your management style?
It is impossible to have a prepared answer for all these kinds of questions. However, taking the time to prepare an answer for some of these will make you better at coming up with good answers for other questions. Additionally, writing down a description of the impression you want to give is also a good way to prepare.
There are 4 main types of technical questions:
- Explaining a previous project
- Case-type questions about how you would approach a certain task
- Theory questions (here are some examples)
- Practical exercises
Again, the range of possible questions here is almost unlimited. If you’re applying for a big company you may find some information about their interviewing style online. Having more experience with data science projects will give you an edge on technical questions but you can always get blindsided with a theory question about an algorithm you’ve never used. If this happens, acknowledge your ignorance and offer another subject about which you could talk.
Finally, when it’s your turn, asking a couple of questions will make you look interested in the job. You can spend 30 minutes googling the company before the interview to stand out from the competition by asking interesting questions.
How I run interviews
I have run data science interviews at both my current and past jobs. One was for consulting, the other for a SaaS that optimizes retail stock management. In both cases, interviews consisted of a data science business case, in which candidates have to solve business problems using analytical tools. It’s not so much about coding (in fact we don’t do live coding) as it is about problem-solving and knowing when to apply each data science technique. More specifically, what I look for when interviewing is:
- Problem-solving, understanding business problems and developing data-driven strategies to solve them.
- Communication skills, capacity to explain complex concepts in a clear and concise manner
- Leadership and initiative, ability and willingness to propose and run projects as well as to mentor more junior colleagues.
- Code craftsmanship, love for writing clear and easy to maintain code, while being conscious of the problems associated with excess complexity.
- Analysis depth, for example by identifying confounding variables and getting to the root cause of issues.
Just keep in mind that this is based on my personal opinions and what my company needs. Other interviews may be different.
How to get better at interviews
Many people struggle with interviews. The good news is, practice can help a lot.
Practice your introduction in front of the mirror until it’s perfect. Create or get some interview scripts, and get a friend or relative to interview you. Or if you’re still in college, maybe get together with some other students to interview each other.
After 5-10 mock interviews, you will be more articulate and more confident in yourself.
If you feel like anxiety is an issue in interviews, try to do breathing exercises to relax before going in.
Tying it up
The job search doesn’t end when you get the offer, but when you sign the contract. Now it’s time to negotiate the terms. If there is something you don’t like, you have a right to say it and try to reach a different agreement. This can range from salary to vacation days.
Words of encouragement
Finally, the most important thing to keep in mind is that getting a job like many things in life is a numbers game. No amount of effort and skill will guarantee that you get the job. They may forget about your CV, the position may be covered by the CEO’s nephew, or the recruiter may be an ass.
Additionally, for every posting, there may be lots of applicants. So don’t lose faith and don’t let rejection bring you down. Even if it’s tedious and it takes a long time in the end it’s still worth it.
Assuming you send 50 applications and each of them takes you about one hour. Then interview with 10 companies, spending an average of 3 hours with each. This would be a total of 80 hours, which is about 4-5% of what someone with a full-time job works on a year. If you get a 10% raise it’s a great investment. If you get 5%, better conditions, or a more fulfilling job it’s still well worth it.
So, go get it!