To PhD or not to PhD?
…That is the question.
Further study often seems the most appealing route to go down, and for many companies out there it’s often heralded as a must have for data science, but it isn’t always so and isn’t always pivotal in advancing your career.
A Ph.D. is your own original, unique research into something not necessarily covered from that angle before. The content of a data science Ph.D. should showcase new findings in the field that will make an impact, or contribute to a particular subject. A lot of people pursuing Ph.D.s are driven by a passion for their area of study, more than thinking of a specific role they can enter into afterward. They can take a long time to complete, so having a particularly keen interest in the subject, and passion in the field will drive your success. Ph.D.s are also a gateway into being published at top conferences in your chosen field, like ICCV (International Conference on Computer Vision), NIPS (Conference on Neural Information Processing Systems) and ECCV (European Conference on Computer Vision) to name just a few, which undoubtedly has a huge impact on a researcher’s career path.
Carrying out and writing about such in-depth research shows your ability to think critically about problems and how to provide solutions; researching new algorithms or improving existing ones. It often shows that you’ll be able to push the boundaries of what a company is doing once you enter into the industry. These are attributes desired by companies as they look to handling growing big data sets and solving even more complex problems.
Not just this, but research gives you an upper hand when it comes to fulfilling the demand for data scientists that can tell a story – i.e communicate at a higher level, their data insights within companies. Having a deeper knowledge of methodologies and the foundations of data science, you’ll typically exhibit a deeper understanding of theories and work using these, rather than applying algorithms from other sources.
Data Scientists holding more specialist Ph.D.s such as computational linguistics, computer vision or Artificial Intelligence can often command a higher salary in relevant industries, as well as the ability to enter a research position in your specialist field.
Not To Ph.D.
You may believe that Ph.D. candidates will always have an advantage when applying for roles, but that is not so.
It all depends on the type of career you want to pursue and the type of projects you want to work on as many data scientists have fruitful careers applying methods and algorithms rather than deeper research. It’s completely plausible to have a successful career in data science without a Ph.D.
Working your way up in a commercial role for example as a data analyst, you get the chance to work with data scientists, from whom you can build on your own skills and experience. With the right focus, you could soon find yourself a Data Scientist down the line. You’ll be more focused on applying current methods to solving problems and also have the opportunity to develop new algorithms.
If you are applying machine learning to a problem, you’ll need working knowledge of algorithms however, you’ll be more concerned with how to use it to provide a solution to what you are working on. With a focus on broader concerns surrounding the data, like its source, validity, how to format it, mining, or analyzing.
A Ph.D. is not necessarily indicative of a candidate’s ability in finding solutions to business problems. Engineering roles for example often prefer industry experience and hands-on coding more suitable than years of further study. If you are a Masters grad and seeking to begin your career rather than undertake a Ph.D., it is definitely worth gaining exposure in a particular industry to kick-start your commercial career.
There is no right or wrong to this, and there are some fantastic data scientists out there who do not have a Ph.D. As with most things, it comes down to personal preference and what type of career you want to pursue. Know your strengths and play to them.
Link: To PhD or not to PhD?