From PhD to Data Scientist:
5 Tips for Making the Transition
Originally posted by Douglas Mason with Valerie Bisharat
Douglas Mason, Harvard Physics PhD, Insight Fellow, and Data Scientist at Twitter, outlines his advice on transitioning from academia to data science.
About a year ago, I began my unexpected but rewarding transition to industry after completing my physics PhD. My dream for years before that had been to work as a physicist in the National Laboratories, but when the time came to do so, that fate just didn’t feel right.
Before Insight, I had virtually zero knowledge of how to score a job at a tech company like Twitter, Facebook or Google. Instead, what I carried with me before, during, and after the program was a relentless enthusiasm. I’m certain that was key to my success, leading to my current role as a data scientist at Twitter.
Here are my top tips on transitioning from academia to the tech industry:
1. Show that you want it. I’m now part of the interview process to select new hires at Twitter. You wouldn’t believe how many people come through here saying, “Well, the academic thing just isn’t working out for me. I guess I’ll do this now.” We can tell on your resume and in the interview if you didn’t put in any effort to look impressive. If that’s the case, what’s going to happen when we hire you? We want people who actively want to be here.
2. Emphasize the parallels between your thesis work and potential professional projects. My data science career is hugely impacted by having written a thesis. Remember that as a PhD you’ve essentially conducted a five-year project that you were totally responsible for. You’re also constantly giving talks, presentations and summaries of your research, which is exactly what my job involves as a data scientist. Work projects are essentially like an entire thesis compressed into one or two quarters. Only now, you have a bunch of colleagues who will help you figure the problems out. In your resume and the interviews, find ways to draw parallels between your work experience from your PhD and the responsibilities outlined in the job description.
3. Take time to practice your hard skills. They aren’t easy.
- Study your algorithms and data structures from a lot of different sources.
- Give yourself time to learn recursive programming — it’s a different way of thinking, so you can’t do it in a night.
- Review your statistics. Know your different regression types, as well asp values and t-tests.
- Know how to calculate expectation values in combinatorics problems.
- Learn and work with SQL.
- If you’re using Matlab, Fortran, etc. it’s time to make the transition to Python or R. Build fun side projects to practice your skills.
4. Interact with the tech community as much as possible. Learn the language, the lingo, how people talk, how people think. Learn what people value. If you go in having done none of this, you will sound like an alien. You have to show, by doing, that you’re willing to learn the vernacular.
5. Choose a company with an employee size that fits your needs. I love that Twitter is a medium sized company — it’s big enough that I can learn a lot from the experts around me, but small enough that I can have a big impact. For me, that’s an ideal balance. Consider your desired mix, and get lots of advice from veterans along the way.
Remember, this process will throw lots of unknowns at you. In fact, the unknown is the only constant on this path. Get comfortable with that, stay focused, be positive and work hard. Going out on a limb is usually worth it.
Interested in transitioning to a career in data engineering?
Find out more about the Insight Data Engineering Fellows Program in New York and Silicon Valley, apply today, or sign up for program updates.
Already a data scientist or engineer?
Find out more about our Advanced Workshops for Data Professionals. Register for two-day workshops in Apache Spark and Data Visualization, or sign up for workshop updates.
