Is PhD about studying something niche
I'm back! This week, I would like to discuss one key reflection on the PhD surveys.
Industry - Are PhD confined to a niche area?
A few months ago, I posted two surveys to get people’s thoughts on why and why not study for a PhD (check out these two posts: Why (people in robotics) study PhD and Why (people in robotics) do not study PhD). At a high level, I agree with most ideas from both sides. However, there is one interesting thing that I noticed: many people do not consider a PhD because they are not interested in diving deep into a niche area before knowing whether they will truly enjoy it. To me, this statement needs some clarification.
First, what is a niche? Is “controls” a niche (even though similar control theories can be applied to very different systems, e.g., motor control and building heat control)? Is “robotics” a niche (even though it can cover areas from mechanical design to machine learning)? Truth is, PhD students often study more than one area during their 5- to 7-year span. Based on my personal observation, no PhD that I know starts and ends at the same “niche”. Everyone enjoys exploring and trying new things. Take me as an example. I started off working on optimization-based motion planning problems. Then, I took a diversion into mechanical design and mechatronics for prosthetics. Next, I began working on sample-based and search-based motion planning. Finally, I explored reinforcement learning. I know people who explored more than I do. And even if they don’t, that is their own choice. There is no magical force that confines a PhD to one niche area.
Second, I have a challenging thought: how does landing an industry job make a person not “dive deep into a niche area”? Applying to a PhD is us taking initiative and choosing the area to dive deep. However, I argue that landing on any job is also diving deep into - well, the day-to-day job, whatever that might be. The tricky part is that now you lose control over what the niche area will be. And most people will agree that arguing to work on a specific area is way easier as a PhD student compared to an employee in industry. Some people get lucky and are able to try many things in their first job. But most will likely stay in one “area” for a while until they switch teams, scope (e.g., promotion), or job.
Third, not every PhD student is entirely sure about the areas they want to work in down the road, and it is likely to change anyway. If you read the “why study” survey, you will see that many are interested in learning how to formulate and solve a problem; these are more general skills that one can take when they move on to a new area.
The point I’m trying to make is that there are no real constraints in either choice. Getting a PhD does not mean you have a narrower career path. Jumping straight into the industry doesn’t mean you cannot learn more and delve deeper into specific areas. However, there is one big difference. Industry is about business, academia is not. This difference makes the two environments distinct, and therefore, the question that one should really ask is: What environment do I want to be in for the next five years?
Here are some observations about the two environments:
Academia promotes learning, exploration, and research. Almost everything is established around this idea. Resources towards these aspects are also abundant. For example, anyone can knock on a professor's door to ask a question, at no cost.
The industry promotes creating value for customers (i.e., generating revenue). Anything that ties with this idea is easy to accept. Technically, people in industry can do everything that people in academia are doing - as long as they can make a business case out of it. In the industry, everything is added with the flavor of business. Personally, I find “learning about business” to be a big part in enabling me to exercise my PhD skills.
Therefore, if our goal is to learn about linear control systems, academia is certainly the best environment to learn. However, if our goal is to make money, then the industry environment will be better because that is the core idea behind it.