Many companies have ‘PhD in e.g. computer science, machine learning, maths or [insert quantitative discipline]’ in their minimum qualifications for data scientists, machine learning scientists or quantitative researchers. Why is this so? Here are a few of their responses:
- PhDs bring real expertise in their areas. We need expertise in cutting-edge methods and technologies and it’s unlikely that people who haven’t spent a lot of time immersed in the topic area would have that knowledge.
- PhDs have experience doing long-term research. People we hire may be given open problems or projects that they tackle for a year at at a time. We usually spend the first month or going in some direction and then realise it’s the wrong direction. If candidates haven’t had previous long-term research experience they might not be able to do this well (adapt and persist).
- PhDs have experience doing research on their own. Even when you do a Masters most of it is taught to you.
The reasons seem to fit into two categories: direct domain expertise and research experience. The ‘research experience’ qualities are more vague and don’t come only with a PhD, but it’s accepted that the probability of PhD students doing well in long-term research is higher because they have done it. So when companies can choose to hire almost exclusively PhDs which have proven research experience, why wouldn’t they? It decreases risk and filters out candidates in hiring, and hiring is costly.
What does this mean? If you don’t have a PhD but want these jobs, do you have evidence that (1) you have real expertise in your area (optional depending on the role and company), (2) you can do long-term research and (3) you can do research on your own? If not, there are many open problems waiting. 🙂