Recently, I got an email from a smart undergraduate student who was interested in research. I think his situation might be a common case so here I added my reply in the hope that this can be helpful to other students. Below is the modified email message.
I have been talking with few graduate recruiters and found out that graduate schools don’t offer scholarship unless I am doing PhD or thesis (which I had no idea what that meant). All I wanted to do was to get a masters degree to gain enough knowledge on AI and deep learning and work for companies like DeepMind or Neuralink to bring AGI and ASI to real life. And now I am overwhelmed and lost with what I’m supposed to do. My options are to loan my way through graduate school (but I can’t do that since I am already on a huge debt for my undergraduate school), or get a PhD (this will take me forever and I will never be able to pay my loans back with the stipend), or find a job (until I am able to pay by loans and fund my graduate school but I really want to be on AI). I recently figured out that my major (computer engineering) wasn’t right for me. I decided to switch to computer science but I needed to take more classes so I am switching to software engineering to graduate on time with right the classes. My mind’s all over the place and I have no idea where to seek help from. I’m only halfway through the book even though I love it. I have been dying to learn Reinforcement Learning but I haven’t got time even to start searching about the resources.
Thank you so much, Dr. Park! I hope you have a wonderful rest of the day.
Below you can find my advice. TLDR, it is okay to get a job if you are on a debt. If you think you are interested in the research, try to study on your own for a year or two. If you are still interested, you can consider a PhD.
My answer to the student
Hi, Tomas (Name changed for privacy)
I understand your position.
Below is my advice. The choice is yours. There is no right or wrong choice. The choice is based on who you are and in return shapes who you will be.
to loan my way through graduate school (but I can’t do that since I am already on a huge debt for my undergraduate school),
The graduate school here must be master’s degree, right. I don’t recommend it, either. You cannot get enough knowledge on AI topic and you don’t have enough time to create a knowledge that will make you attractable hire for companies like Deep Mind.
or get a PhD (this will take me forever and I will never be able to pay my loans back with the stipend),
It does not take you forever. It is actually somewhat short to do what I describe above. It will pay you back reasonably well if you are successful. But the decision cannot be made solely on financial reward because this financial reward is probabilistic. Better motivation would be the pursuit of knowledge or curiosity. And the reward is auxiliary. For example, you usually do not pursue professional baseball career because most of baseball related careers are not well-paid. Only winner takes it all.(I hope the chance is higher in the CS.) But many still pursue it because they enjoy baseball anyway.
or find a job (until I am able to pay by loans and fund my graduate school but I really want to be on AI).
This is an actually good option. Get a job. Get decent salary and pay your loans. The trick is that you can study AI on your own with online learning education resource nowadays. This can be sometimes better than what the Master’s degree can offer. When you study enough and you have a non-trivial question, it will be a good time to pursue PhD. Of course, consult with professors about your question.
Attending academic conference is a good investment. Studying alone is actually harder than you might think. You will be busy for the job (actually everyone is busy, right? Even my 9th grade daughter is busy because she has to watch three hours of youtube videos in addition to all the school work and violin and archery practice) and you may find the AI study is not so fun. The good news is that if you don’t like it, you don’t have to do it. It means that you are probably not good fit for the PhD. Don’t worry. You have a job already and you can focus on the software engineering career which can be also very rewarding. According to my opinion, I think about 5% of the people with good school grades are actually good fit for the PhD. The main reason is that we get good grades by learning what others have found. But PhD is about finding new knowledge which is very different from learning what others have found.
As a summary, I think you have a potential. If you really want to study AI and deep learning now, you can get a PhD now. If you are not sure, it is okay to get a SWE job. If you find that you keep studying on AI and DL for one or two years on your own, you might consider getting a PhD.
Regards
Deokgun