The world’s leading publication for data science, AI, and ML professionals.

Transitioning from Pre-Health to a Career in Data Science/Tech

Tips/Resources for making the career change in your Undergraduate

Image from Unsplash
Image from Unsplash

After spending two years as a pre-medical student at the University of Virginia I slowly realized I could not and did not want to sustain the level of dedication and time required for the next 10 years of my life (I was on pace for a lot of gap years lol). I transitioned slowly into a Statistics major and decided I was interested in a career track called "Data Science". I didn’t really know what the track entailed nor what it even meant, I just heard the words being thrown around a lot as the new lucrative job of the century. Eventually I came to understand the nuances behind the field and slowly fell in love with the theory and applications of Data Science/ML (I will use both interchangeably in this article) in the real-world. However my learning curve and transition was steepened due to some mistakes I made as I ventured into Data Science and Tech in general. I wanted to share some tips and advice to help ease and guide this transition for whoever is looking to switch into a career in Tech. There’s so many buzz words and an ever expanding skillset that can seem super intimidating at first, but with time and a proper learning path everything becomes familiar. I’ve also attached resources that I found handy in my transition throughout the article.

Become a Programmer First

This applies especially to those trying to venture into Data Science, it’s obviously a skill for those trying to go directly into Software Engineering in the Tech realm. During my time as a pre-med student my course content and studying was very memory-centric. Pre-med comes with an enormous amount of details about the most minute topics and my general approach was to always memorize first and apply later. This mindset made programming and Computer Science very hard for me to stomach initially. I was intimidated and scared at first by the hundreds of lines of code my Computer Science friends had and I looked for any way around having to program. This was my first big mistake in my transition. I focused on theory and mathematical portions behind ML models which intrigued me and were much more appealing to me. While this was great in building my theoretical knowledge base behind various ML algorithms, I couldn’t really apply it. Without concrete programming skills I struggled to build any ML powered applications or models. It wasn’t until I sat down and picked a language (Python) and dove deep into Object Oriented Programming and Data Structures/Algorithms that I was able to gain a level of confidence in my programming skills. The number of languages and frameworks seem incredibly vast and alien to non-programmers at first but after developing the confidence and intuition in one language it is truly not difficult to pick up these other tools. My ability to showcase my ML skillset vastly improved as I improved as a programmer, I was able to see the actual effect of ML in real-world applications. I’ve linked a course I used to start my learning journey with Python, which is the most common/popular language for Data Science. In terms of online courses such as this in general, I don’t recommend being hell-bent on finishing the course but rather getting the fundamentals then practicing problems or projects on your own. If you don’t want to pay for a Udemy/Coursera course, sentdex is also an amazing YouTuber who specializes in Python and ML. At the end of the day, there are limitless resources on the Internet and whether it’s through that or coursework I strongly recommend gaining familiarity with programming fundamentals before delving into the world of Data Science.

Changing Mindsets

I touched on this slightly before in the previous section, but one of the greatest changes in this transition was how I attacked problems. I had to slowly shift from a memorizing heavy approach to looking at problems more intuitively. At first I almost tried memorizing syntax because of how my mind had been programmed to think. After realizing how unfeasible this mindset was in the field of Computer Science (and after discovering Stack Overflow), I came to understand that programming was something to be mastered through repetitions. Whatever language I was learning was just a tool to solve a logical problem and stitch together solutions for a bigger application. The shift in this mindset took a few months to come as I became more comfortable with myself as a programmer. It also initially helped to draw analogies to my science background to help decrease my unfamiliarity with the new field. For example, with Object Oriented Programming I strongly associated with the field of Taxonomy to understand fundamental concepts such as Inheritance and more. I also didn’t abandon the memory based approach entirely as I often looked for patterns in code and it proved extremely useful when it came to optimization.

The Machine Learning Part

So the part that "matters", the part that truly composes of what "Data Science" is. In reality much of Data Science is not the model-building portion, which is why you need strong programming fundamentals to properly prepare/cleanse data and understand deployment. For the actual theory behind ML, I strongly recommend getting a foundational understanding of Probability and Linear Algebra. If you have the time while in school definitely try to take some coursework in both disciplines. If you’re more of the learn on your own type, there’s many resources online once again. The Khan Academy learning path for both topics is a good introduction to the field. Probability wise, make sure you have a strong Calculus background in particular. In parallel with this learning or after I strongly suggest taking an Introductory Level ML Class. Once again for online learners I recommend Jose Portilla’s ML Bootcamp. He’s a great instructor to get you running off the ground and offers a good introduction to the various ML models. Not everyone has to be super theoretically polished and he’s a very good resource for those seeking to be more application heavy. For more theory centric courses which require greater Mathematical and ML knowledge, I strongly recommend exploring LazyProgrammer’s course offerings. Regardless of the course you take to starting ML, much like programming you only get familiar with real-world datasets through application and repetitions of practice. I recommend starting with Kaggle to explore some untidy datasets and then try scraping/collecting your own data to build ML powered applications for personal projects or ideas that you have.

— – – – – –

Much like the fields of Data Science and Tech, this article may have seemed like a lot dumped on you at once. The way to learn in these fields is getting your feet dirty and diving right into practice problems after getting an introduction. Getting over that initial block and aversion to something new is common for many Life changes. Embrace the change and the opportunities that will come with this field. Your previous experiences enriched your mindset and make you more diverse if anything. To summarize the general learning path I’d recommend for those trying to make this transition is as follows. Become a programmer (master one language, OOP & Data Structure), develop strong mathematical foundations in Probability and Linear Algebra, understand the various ML algorithms/models at an application/theoretical view point. After and even during these three general steps: practice, practice, practice. The beauty behind Data Science and ML is the ability to apply the field to any real-world problem you seek to solve. If you’re still interested in pre-medical problems, there’s so much scope for ML in that field. Whether it’s image classification of X-rays or using NLP to analyze medical reports there’s so many interesting projects to tackle.

I hope this advice has been useful for anyone who’s made it this far in the article. These were just some of the tips that helped me make my transition into the technical world. Feel free to connect with me on Linkedln or follow me on Medium for more of my writing. Share any thoughts or feedback, thank you for reading!


Related Articles