Applying for a data science job can be intimidating, particularly if it’s your first and you don’t have any prior experience. While there are plenty of jobs out there for data scientists, the nature of the modern job market is still such that you’re likely competing against dozens of other candidates, and potentially hundreds or even thousands for high-profile jobs. Hiring managers may make a decision after just thirty seconds looking at your application. What can you do to stand out?
To find out, we spoke with Refael "Rafi" Zikavashvili, the CEO and co-founder of Pramp, an online job interview practice site that offers free practice data science interviews. That’s a great resource in its own right, but the nature of Rafi’s work is such that he’s spent a lot of time both thinking about and actually doing data science hiring. And he’s got one big tip for data science job applicants who want to stand out: kill the cover letter, and replace it with something better.
Better Than a Cover Letter
“Cover letters are antiquated,” Rafi says. “At least in tech, you don't really see them often and frankly, it's not a practice I would recommend anyone to do.”
“But there is something similar to a cover letter that, in my opinion, is much better,” he says.
Specifically, Rafi suggests that you send potential employers an email pointing out a challenge that their business is likely facing and how you would solve it. “That will super impress me,” he says. “Somebody who can actually identify problems just based on public information, and more than that will go the extra step and actually suggest a way to solve them? That will blow me away.”
“That person gets an instant interview from me, basically. No questions asked. I would even skip the resume at that point.”
In fact, Rafi says, this doesn’t even necessarily need to be sent in connection with a formal job application. “It could be a candidate that really likes the company and just decides to check if there is anything available for them, or in response to an open position,” Rafi says. “It doesn't matter. What matters is really what the action that they take. Being proactive about helping the company, that's something that always impresses me.”
Of course, this isn’t something you can do for every job you apply for if you’re carpet-bombing the market and applying to every job posting you come across. Because it requires a significant time investment, this is a technique best saved for the job postings or companies you really care about.
“Find a company that you really like, that you're passionate about,” Rafi suggests. “Analyze that company. Find a challenge that you think you can help with. Come up with a solution. That's the best advice that I can give somebody. That's a way to stand out.”
In other words: stand out by showing that you actually care about the company in question, and are interested in and engaged with its business problems.
High Stakes: Doing Data Analysis as a Job Application
First, you need to select the company and/or job you’re interested in very carefully. This is a time-consuming process, so you should only undertake it for positions that you’re very, very interested in. Understand that this is a high-risk play, so be sure that the potential reward—the job—is worth it for you.
Second, you need to establish that you have a way of contacting the hiring manager, CEO, or some other decision-maker. It’s likely that the job application process obfuscates this information, but if you’re going to all the trouble of doing a custom analysis, you want to make sure that somebody actually sees it, so sending it to a human’s email address is better. The good news is that finding this shouldn’t be too hard if you know the person’s name; services like Hunter.io and VoilaNorbert can help you find business emails with a high degree of certainty if you know the person’s name and company URL.
Once you know the job is worth it and you’ve got a reliable way of contacting a real person (hiring manager, CEO, Data Science lead, etc.) at the company, it’s time to start your research. You may be surprised just how much public market and even company data is out there if you do a little digging.
What you do will be highly dependent on the specifics of the company you’re looking at, but you should think about this the same way you might think about any business-focused data science project, following steps like:
- Identifying business problems the company has or is likely to have
- Identifying what kinds of data might be useful in suggesting solutions to these problems
- Collecting, cleaning, and organizing the data
- Doing some exploratory data analysis and visualization
- Drilling down into the most fruitful areas for further analysis
- Visualizing and/or writing up your analysis
If you need help with any of the technical aspects of this kind of project, we’ve got easy-to-understand, hands-on online courses that’ll teach you everything you need to know.
But assuming that you’ve got all the technical skills down already, what’s most important here is that final step: communicating your analysis clearly in a business context. In other words: What’s the problem? What’s the solution to this problem, and why? That’s the most important message you need to get across. Don’t throw a bunch of advanced math or code at a recruiter, even if you think they’ll understand them.
The point of going to all this effort is to demonstrate that:
- You’re interested in this company and already thinking about its business problems on a deep level.
- You have the technical skills needed to analyze business problems at this company
- You’re proactive, and a self-starter
Make no mistake, this isn’t always going to work. Hours and hours of work could end with an email that just gets you a polite “no thanks,” or simply no reply at all. But if there’s a company you really like and you’re concerned about standing out in a sea of applicants, this outside-the-box strategy could be what you need to get yourself to the top of the hiring manager’s list.
Lower Stakes: Using This Advice in Traditional Applications and Interviews
The basic idea behind this is just that companies want to see you’re thinking about and grappling with their business problems. If you don’t want to go all-out with a company-specific data analysis project, there are still ways that you can help yourself stand out by demonstrating that you’re doing this in the context of a more traditional application process.
If you’re applying to a few different companies in the same field, for example, you can do a data science project that has some relevance to that industry and then include that in your application. This won’t have the same standout impact as doing a company-specific project, but it does demonstrate that you’ve done some thinking about the kind of work you’d be doing in the industry of the companies to which you’re applying.
Once you’ve made it to the interview phase, you’ll have another opportunity to show the extent to which you’re thinking about the company and its business problems. Asking the right questions of your prospective new employer can really help you stand out. From a great data science applicant, Rafi says, “I would expect some very probing questions about some of the flows that we have, and some of the assumptions that we have, the KPIs that matter to us.”
“If I don't see a data scientist asking those questions, that's a red flag for me,” he says.