Should I go to grad school? I'm interested in ML & AI and wish to work in this field. To give some background, I'm a third year undergraduate at a research-heavy university. I interned at Salesforce this summer at a machine learning team, and will be joining Facebook for next summer internship. Going to grad school, pros: - Work on more interesting projects when I join industry later (research labs only seem to be taking PhD / Masters) - Possibly higher pay (I've heard stories of research scientists at top firms being paid 7 figures) - Deep learning is all the hype right now - I'm still young and early in the game and free to pursure opportunities that I may not have the chance to later in life (eg: when I have a family) Going to grad school, cons: - Already a lot of engineering teams are using ML in their work (although it won't be state-of-the-art — I've heard most production teams even at the big 3 are still sticking to pipped-up random forest & boosting models) - Money & time commitment - I'm getting pretty sick of school (maintaining GPA is stressful) - Deep learning could be a fad and specializing in this area might not have negative affects in terms of future career opportunities Are there any engineers in the industry currently regretting going to / not going to graduate school?
By the time you get done, ML will be yesterday’s news. There are also a limited number of current problems that ML can effectively address, it requires huge amounts of data and a clean signal to predict on. The world doesn’t really need that many good ML researchers, it mostly needs code monkeys
Thanks for the reply! Are you working in a ML team at Google by any chance? How fast is the adoption of latest research to production there?
Good time to go to grad school is right now !! I know people who want to do a Phd at 26 or 30/35 years of age and I don't think they can go back to a poor pay life. Also your social life goes for a toss. Hard to manage with family or kids! If you want to do a phd later on in life - Then better idea is to land a job in a research organization e.g. MSR or FAIR or AML - Some of the biggest skills you learn during a PhD : 1. An appreciation and ability to separate cutting edge work from incremental work a.k.a low-hamging fruits. Many fresh grads at FAANG work on hacky or low-hanging fruits. Quick results but not focusing on long term impact. 2. Ability to come up with high impact ideas esp. for an ambiguous or tough problem - Something no one has thought of before. 3. Ability to go deep (really deep - Like super deep-dive using Amazons LP terminology ) into one tiny area of technical expertise. This deep dive can be at all levels - Math, experimentation, design, etc. 4. Ability to sell your work scientifically through publishing papers. If you are getting many of these in your current role or a future role - And you are not at a stage of life where PhD is an option - Then forget about it! There is more to life than slogging 15hrs a day in a lab (And I can say that because I have gone through it).
Grad school is hawt. If you go, I'll be inclined to deep-learn your bod ;)