You finally finished that college CS degree, completed hundreds of hours of training online, got your certifications, or simply have the skills you need to transition. The bottom line is, you are ready to start working in data science, an incredibly exciting field.

And then, you start looking for a job, and all openings ask you for 2 to 5 years of experience working in data science. Wait, if all jobs ask for experience, and there is no way to get experience unless you get a job, how can you ever get out of this impossible-to-solve loop? Well, it is not impossible to solve and there are a few options available.

Option 1: Get an Internship.

This has been the go-to option for many new professionals of all areas, especially since the genesis of unpaid internships. Becoming an intern can be an amazing experience and get you started on the right path to an incredible career in data science. The only issue with this is that the good ones are scarce, and many offerings are not worth the effort. If you do decide to go take this route, make sure that the internship is with a company that has an established data science team, that way you will learn from people with experience. Also make sure you will be doing actual work, not just menial tasks (and please remember, in data science, cleaning data is NOT menial, it is VITAL).

Option 2: Project Building

As you have heard over and over on the internet and from other data scientists, building projects is a great way to learn, get some experience and showcase your skills. The biggest issue with this approach is that many people build great projects that, although interesting, have no real-world applications. In other words, you get to practice the theory you have learned in a theoretical background. Don’t get me wrong, I love projects, they are the best way to learn in my opinion, but programming another time series for the stock market or another COVID tracker will bring you very little attention and actual experience. If you decide to opt for this path, make sure you look for an interesting project, something that you are passionate about, and hopefully, that has real-world applications. Also, be sure to have your GitHub account updated with all the info on your project and use social media to showcase it every step of the way.

Option 3: Competitions

Kaggle, and other similar places, are filled with interesting competitions you can join and use to put your data science skills to the test. building a better model, and you might even get some money out of it if your model is good enough. Yet, there is a quintessential problem with these competitions, as they focus mostly on the model building and not on any other part of real-world workflows. For example, you get a predefined problem to solve and a clean dataset. This NEVER happens out in the wild. You need to figure out what type of problem the business or organization needs to solve, then figure out how to get the data. If it exists, you will need to clean and organize it, if not, you need to find it or come up with ways to get it. Then and only then can you start building a model, and once you do, you still need to work with development and engineering teams to figure out how to deploy your model into something useful. Competitions will test your model-building skills but little else. Yet, if you do want to put your skills to the test and become a Master or Grandmaster competitor, my only advice is to put together a team and compete as such, this way you will develop teamwork skills that are vital for your job success in the future.

Option 4: DATApreneurship (THE BEST OPTION, and yes, I am biased)

Finally, we get to my preferred option to start working in data science. Why? Because as you will see, it combines all of the previous options and adds quite a bit to your experience. DATApreneurship means starting your own Data Centered business where you are the founder, the boss, the Data Scientist, and the intern, all rolled up into one. You get to decide an actual problem to solve, that has real-world application and commercial value (in other words, something you can sell to others). Then you get to put a team of like-minded individuals together, whether other data scientists, or better, people from other backgrounds and other skills that will add to your ideas. Then it is time to find how and where to get the data you need, you also get to clean and organize it, do all kinds of EDA (Exploratory Data Analysis), come up with a hypothesis, test them, use what you find to build your models (and please, always start with the simplest algorithms), figure out how to deploy them, and, last and yet most importantly, go out, and sell.

With DATApreneurship, if all goes very well, you will have literally built your own job and your own company at the same time. And if, like many entrepreneurial attempts, you fail once, twice, or many times over, you will get the experience you need, you will be able to showcase your skills in solving real business-oriented and useful products and services. Every day you work on your DATApreneurship idea, is a day you gain actual work experience.

Conclusion on Working in Data Science

Even if you are looking for a job, never stay still. Start working on your own ideas, on your own projects, from day one. Never ever wait for a job to start working.

Join me at ODSC Europe’s Virtual AI Career Expo to talk about the advantages and steps you can take to start Building your Own Job and becoming a DATApreneur.


About the author/ODSC Europe 2021 speaker: Jack Raifer Baruch

Head of Data Science at ADA Intelligence

Data Scientist and Behavioral specialist working on predicting future behavior through the modeling of weak behavioral signals, and building products and services to help people improve themselves and their future outcomes.

DATApreneurs on Discord:​