Seven Tips for Crafting a Great Data Science Resume

Helping you create an excellent, brief resume for your next data science job

Hello there friends! If you’ve paid attention to how the data science industry has moved recently, you might notice that there’s been a lot of shift in employment recently. With the advent of many companies going full virtual in a post-COVID world, many people are reassessing their current employment for more preferential opportunities. (I hesitate to use the word “better” opportunities because “better” is relative.)

While I am not in the job market myself, I thought I would write this post for all the people considering a change in data science employment. In my opinion, most resumes aren’t very good because they tend to adhere toward outdated “best practices.” The evolution of technology has changed things in such a way that places like LinkedIn allow a person to “word vomit” their whole job history, which means that’s not exactly necessary out of a resume anymore.

Moreover, I don’t think old resume tips captured this one reality well at all: human beings are highly irrational. We all are irrational at least in some area of our lives. While we’d like to think we are irrational when it comes to selecting an ideal candidate for a job position, irrational factors tend to seep in subconsciously. Because of this, a number of these tips are written to account for these irrational factor. To be completely honest, there is at least one tip in here that I personally find completely ridiculous. You learn which one that is in a bit.

Before we jump into these tips, I updated my own personal resume to help with a concrete example of how adopting these tips might look like. Aside from being a machine learning engineer, I also do some graphic design on the side and used Affinity Designer to create my resume. Affinity Designer is great, but like its more infamous counterpart — Adobe Illustrator — it’s a tool designed for more professional graphic design usage. I personally think it’s worth considering you considering using this software as being able to create basic graphic design artwork is super, super handy. (For example, all my blog title cards are created in Affinity Designer.) But if you want to keep things simple, I might recommend a platform like Canva or EnhanCV.

Okay, let’s jump into our tips!

Created by the author

Thinking back on an earlier comment about how technology has shifted how people are hired, this is especially true for data science candidates. If you apply for a position today, chances are that there is a form-like application process or something that scrapes information from your LinkedIn profile in addition to you submitting a PDF-like resume. Moreover, an employer is most likely going to assess your skills via something like a coding exercise, so it’s not exactly necessary anymore to ensure you get every little thing about your skills and experience on your resume itself. Chances are that the employer is sifting through a whole stack of these resumes, so you want to keep things as brief as possible. There is no reason your resume should have to go longer than one page. And if you arrange things like I did, you can get a lot of information on one page!

Consider the traditional, average resume. Most traditional resumes simply include black text on white paper. Now imagine if you were to hang the traditional resumes of all applicants on a wall and randomly include a resume designed like mine amongst them. Standing 30 feet away, could you pick out which resume is mine? The answer of course is yes! At 30 feet away, there’s no way you could read any of the text on anybody’s resume (unless you have eagle-like vision), but you can get a general flair of a person’s design. This is where we start to see human irrationality creep in. If you gave a hiring manager or recruiter a stack of resumes, they’re inclined to fly through the simple, old-style resumes even if they represent highly qualified candidates. They’re more inclined to stop at a specially designed resume just because… well, it looks nice. Again, it’s not rational, but because I want to see you succeed, I’d be doing you a disservice if I didn’t mention this. Speaking of irrationality…

Remember when I said in the introduction that there was one tip in here that I think is ridiculous? This is the exact tip I was talking about. Again pointing to the irrationality factor, people unfairly judge people based on their physical appearance. You’ll notice that I do include my own self portrait, and that’s only because I think my photographer friend did an absolute out-of-this-world job on my headshot. Generally speaking, I would not consider myself a “conventionally attractive” person, so if I had just taken a selfie with my smartphone, I’m not sure I would have included it. Moreover, I’m fortunate in that I am a younger person, and even though this is totally illegal, I’m sure there are some employers that would gloss over a person’s resume if they appeared to be older. Again, the fact that this is even an issue frustrates me to no end, but I can’t change the minds of hiring people. I can only help you, friends, craft your resume in our irrational world.

As we touched on earlier, the resume is really just one piece of a full group of things that comprise what an employer looks at when assessing potential candidates. Chances are that in the form-like application that the employer will ask for things like a link to your LinkedIn profile, but in case it doesn’t make sure that is reflected on your resume. For data science practitioners in particular, I would ensure to include the following things:

  • Your GitHub profile: Most technical employers are familiar with GitHub and the value it can provide for a candidate to showcase their work in the form of code. I’m not going to cover it in this post, but be sure that your GitHub profile is organized in such a way that your portfolio clearly reflects what you would like the employer to see.
  • Personal website: I’m honestly a little bit wary of this one. I personally think that as long as you have solid GitHub and LinkedIn profiles, then a personal website is not needed. What I would recommend, however, is that you get your own domain name and at the very least have it point to something like a LinkTree. This requires a minimal amount of effort and allows an employer to seamlessly navigate around to whatever you would like them to see.
  • Other data science things: I’m thinking things here like Kaggle competitions or published posts on things like Towards Data Science. As somebody who has experience interviewing machine learning engineer candidates for a Fortune 50 company, I can tell you that I personally have never interviewed anybody that has displayed these extra things, and that’s okay. I don’t necessarily expect to see these, but I can tell you that they would definitely be icing on the cake!

This is a super important tip for those just breaking into the data science field. As a mentor for a data science coding bootcamp, I’ve witnessed how students coming out of there struggle to know what experience to list on the resume. I think it is important to put at least one or two previous positions on there just to show that you have general experience in the workplace, but if your last 8 jobs were all at things like restaurants, I wouldn’t list them all to save room on your resume. With that extra room on your resume, I would instead include highlight one or two personal data science projects. When you describe that project, be sure to include the following things:

  • A brief description about the project: Two or three sentences about the overall goal of the project
  • How the project was successful: For my data science bootcamp mentees, the projects they do generally seek to have some sort of positive impact for the world, whether it be for a company directly or in a more “open source” capacity. I definitely encourage highlighting the successes of the project as it helps the employer to understand that the project was worth pursuing in the data science field.
  • The technologies used to support the project: This can simply be a list of the things you used, including Python libraries, open source tools, cloud platforms, and more. For example, you might list something simply like this: “Technologies used: Pandas, Scikit-Learn, FastAPI, AWS SageMaker”

This is particularly important because data science positions can be quite nebulous. Some positions are closer to being general data analysts, where the position focuses more on data gathering and data cleansing. Other positions may focus more on only building predictive models. Others like my own machine learning engineer position may look for more of a software engineering background. Whatever the case is, you will likely need to tweak your resume a little bit to best suit the needs of the company. Here are some of the skills and attributes I might highlight for different categories of data science positions:

  • Data analyst / data engineer: SQL, Pandas, feature engineering, Spark (PySpark), Hadoop, NoSQL databases
  • Data scientist: Applied statistics, predictive modeling, Pandas, Scikit-Learn, PyTorch, Tensorflow, Jupyter
  • Machine learning engineer: CI/CD (e.g. GitLab, Jenkins), FastAPI / Flask, Docker, Kubernetes, AWS, Terraform

Most data science positions have a multi-step process when it comes to hiring potential applicants. These steps generally include some sort of coding exercise, one or more phone interviews, and generally culminate in an “in-person” interview. (“In-person” in quotes since most final interviews have now gone virtual, particularly since COVID-19.) If a candidate makes it to a final round of interviews, they are generally equally as qualified as other candidates who also made it to the final round. In other words, the employer generally can’t go wrong with hiring any candidate that makes it to the final round.

This is where irrationality steps back into play. If two people are equally qualified, how do you determine who you want to hire? If you’ve ever heard of the “airport rule”, that rule basically states this: if you had to be stuck in an airport with somebody for a long amount of time, who would you rather spend that time with? This is why I think sharing personality is important because for this final round of interviews, this can genuinely be the differentiating factor that gets you the job. Again, it’s not exactly rational, but it’s the way the world works. And I’m less frustrated with this one as I was with the headshot one because personality can show how well a person will work with a team, which is indeed an important factor.