14 ms·

I think CS PhD programs in particular can be a little strange. I decided to go back to school to pursue a CS PhD after chatting with a retired mathematician abo

I think CS PhD programs in particular can be a little strange. I decided to go back to school to pursue a CS PhD after chatting with a retired mathematician about his PhD. He was talking about how he took classes for two years to explore different areas of mathematics and to pass the general exams, then he decided on a branch of mathematics (topology) and approached a professor in that area to be his advisor. After working through a few textbooks together, this person posed a question to his advisor, spent a few more years working to answer the question, eventually answered it and published the result in his dissertation. This seemed to be what a PhD was all about, but it's not what I've experienced. I've just completed the first year of grad school, and while the classes have been interesting, the expectation is that you already have an advisor when you arrive (or at the latest at the end of your first year) and you hit the ground running pumping out papers. There seems to be little opportunity to explore computer science as a whole or to work towards one singular result. I recently had the pleasure of chatting with Peter Sarnak about CS graduate education. He said that CS is in a weird spot between traditional engineering disciplines and the liberal arts, and that the need for grant funding causes departments to encourage students to work on whatever projects the professors are leading. But even he seemed a little surprised when I told him that our department's general exam is basically a presentation on the research you've accomplished so far. I think this is unfortunate. First, because it goes against what I wanted out of a PhD. But also because it makes the PhD not much different than working in industry. If in either case I have to write code for some boss and satisfy my curiosities on the nights and weekends, why would I choose to do it for much less money and worse career prospects? I imagine this is causing the best and brightest to avoid academia in computer science, which seems bad for research and technological progress in general.

gms7777 8y ago
I think different schools have different philosophies as far as that goes. There are schools that admit to the department and have students do very little research for their first two years, and there are schools that admit by advisor and you start research immediately. My PhD was the latter, but for me it was actually one of the major selling points for me on the program. I had a pretty good idea of the area that I wanted to work in, and I was always bad at learning from classes as opposed to hands-on self-driven learning. By the end of two years, I'd already gotten a lot of research done and moving towards the end of my PhD I was able to branch out within my area and take my own path in a way I doubt I would have been able to do in a different style program, where I wouldn't have had the same depth of knowledge in my field. All that to say, I think these are two different philosophies in designing a grad program, and each works better for a different style of student. To those that are still thinking about applying, its something that I think should be a serious consideration when looking at program fit.
- throwawaymath 8y ago
  With regard to your first paragraph, I don't think there are any American universities that admit students by advisor. As far as I'm aware, your admission is a decision made solely by the graduate admissions committee. Even if you have an advisor ready to take you on the most they can do is recommend you to the committee, not supplant it. When you're admitted there are still criteria aside from research to work on. I think this is a fundamental difference between American and European universities. In European universities, you can hypothetically join a PhD program without any prior degree at all if an advisor chooses to take you on personally. No such sidestepping can happen in an American institution, and regardless you'll have to take some courses (even if you're only auditing them in effect) before you can begin research. Unsurprisingly, American PhD programs are longer.
  - overeater 8y ago
    Sorry to be blunt throwawaymath, but that's just wrong. I am very familiar with several American universities (including Ivy League) where the advisor can admit directly.
    - throwawaymath 8y ago
      I don't mind you being blunt at all, but do you mind naming those schools specifically? I'm still skeptical.
      - csa 8y ago
        Not OP, but Harvard, Yale, and Penn are three that I can speak to with firsthand experience. I will also say that, depending on the specifics of the applicant, you might be shocked by what admissions standards can be bent/broken based on the desire of a faculty member to have a specific student. It’s not common, but it happens.
  - davmre 8y ago
    I've been on the PhD admissions committee for the Berkeley CS PhD program. Students in many areas are, for all practical purposes, admitted by an advisor. It's not a formal arrangement in the same sense as European programs, but no one gets admitted unless at least one professor will commit to being willing to advise and fund you. For superstar students this is rarely an issue and you have your pick of advisors, but for many admits there is exactly one professor who's identified your background and interests as a good fit, so you're essentially admitted with the expectation you'll work with that professor. Nothing you say is wrong; you're correct that everyone has to do course requirements. But culturally the course requirements are treated as a distraction from research, and students are expected to start working primarily on research from their first semester. I think this is more or less the case in all top American CS PhD programs, and there are reasons to set things up this way, but it's a very different environment from PhDs in most other fields.
chrisseaton 8y ago
What you're describing as what you didn't expect is how PhDs were done traditionally. It's the more recent US model to have classes and exams. I posed a question to my advisor in order to get onto the PhD, started researching immediately, and my 'general exam' half-way through was a presentation and viva. Obviously the question changed totally by the time I finished, but that's the nature of research and where it takes you. I'm not sure why you'd expect to explore computer science as a whole on a PhD. The point of a PhD is to become an expert in one part of the field and to push it to expand what we know - not to gain general wide knowledge. I don't think that was ever the idea - that's what a bachelors is for.
- awelkie 8y ago
  I think it depends on the field. In mathematics I think the expectation is that students need to study for a few years to even understand what the open questions are, and only then can they begin the research. I'm fully aware of what the point of a PhD is. I wasn't expecting to become a generalist, but I was expecting some time to study some recent results and come up with a tractable research proposal. Indeed, if the purpose of a PhD is to expand the totality of human knowledge, why was it expected that I would publish 3-4 conference papers to graduate? Wouldn't one significant result have been enough?
  - cvwright 8y ago
    Often one "significant result" has to be serialized into 3-4 conference papers in order for it to be digestible and reviewable by the referees. There are exceptions of course, like Gentry's fully homomorphic encryption, but those are increasingly rare.
  - gajjanag 8y ago
    > Indeed, if the purpose of a PhD is to expand the totality of human knowledge, why was it expected that I would publish 3-4 conference papers to graduate? This is a question that bothers me quite a bit with respect to EECS (electrical engineering and computer science). In mathematics, one's PhD thesis can often serve as one's first paper (once revised). Funnily enough, the same was true for EECS even at the top tier places in the 1980's - I know of a case at Caltech where the student did not have any publications prior to graduation, and yet went on to become a successful professor. Is it really true that humanity is producing more knowledge in EECS today per student than it was in the 1980's? I seriously doubt this. In my experience, instead what has happened is that the average "delta" in a conference proceeding has gone down significantly, in spite of the vast increase of words like "novel" and "new" being used, and the ballooning of the average paper length. In mathematics on the other hand, top quality authors freely admit that a lot of their work can be "implicitly" traced back to the "big names" of the past; indeed this has to be done for conceptual and historical clarity. I see that you mention Peter Sarnak in your top comment. Peter Sarnak has a a wonderful article at the very end of "The Princeton Companion to Mathematics" on his advice for students. Among other things, he stresses the importance of the "history of ideas", and how it often brings clarity to an entire field.
nbeleski 8y ago
Too many CS postgraduate programs just feel like engineering ones. This is why chosing a proper advisor is so important as your intentions should be aligned. Even then, there are things you might need to do to satisfy your program requirements that aren't exactly what you want to do. Anecdoctal experience: I was offered a somewhat decent scholarship for a Master's program backed by an Oil company, but I would work on specific projects chosen by the company. I sincerely prefered to go for a program/advisor that fit my goal (more on the theoretical CS research) and no grant (the program is free here). This meant that I have to work to pay my bills and study on 'free time'. This works for me personally because I make a more than any program scholarship would pay me while acquiring some industry experience. This can work while I am doing my masters, but It is borderline impossible for a PhD program.
pgbovine 8y ago
i totally know how you feel since i have felt that way in the past as well :) here's something relevant: http://pgbovine.net/practical-reason-to-pursue-PhD.htm http://pgbovine.net/practical-reason-to-pursue-PhD.htm my two cents: find something where the grants align with your interests, or find a brand-new faculty member who has startup money so isn't as bound by grants in the near term. and if things don't work out, simply leave and go to industry -- with a marketable skill set like what you get from being a CS major, there's no way for you to lose. go wherever has work that you like more.
mathattack 8y ago
I think the two subjects are very different. Math has a much longer history. To get to the edge of knowledge you need to know a lot more because people have built proofs upon proofs for much longer. You need an undergrad plus a year or two of grad to get both the depth a breadth of what’s been done in the past. In CS the edge of the field isn’t as far away. New knowledge invalidates more of the old, and you don’t even need a CS undergrad degree to be near some of the edges.
grigjd3 8y ago
It's pretty rare that PhD programs let you truly explore what you want to do. PIs (generally profs) get the vast majority of outside funding so you pretty much have to find a way to align yourself with a prof from day one.
chuckkir 8y ago
In general a professor is an entrepreneur. They are running a small business where they sell their product ("knowledge") to customers (business and industry) by means of sales and marketing (conferences and papers). Grad students are employees of this business, and of course the prof's incentive is to keep the good ones and fire (graduate) the bad ones, thus the reward for the student working hard is more hard work.
- internetman55 8y ago
  But no good student who approaches the grad school process in a sane manner is going to work for a professor whose students stay there for 12 years and don't graduate. A common thing for deciding if you want to work with someone is seeing how long their students took to graduate, where they ended up, emailing them, etc.
- Fomite 8y ago
  Two notes: 1) I am evaluated on the progress of my students. While bad ones will actually be fired, good ones are more useful to me if they go off and get postdocs etc. Some of the most powerful "businesses" in the field are created by the network of people who all trained with the same PI and then went on to found their own labs. 2) A burnt out, trapped graduate student is not actually one that produces good work, in my experience. 3) Several grants, which are how I fund my lab, will look very poorly on trainees making no progress. There's very little exploitation that's worth a program officer at the NSF wondering if I'm worth there time. Are there bad, exploitative PIs? Absolutely. But these people are assholes, not the only logical outcome of the system.
gervase 8y ago
> But also because it makes the PhD not much different than working in industry. I believe this is viewed as a feature, and not a bug. For the most part, academia serves as one of the few reliable ways to emigrate to the U.S., and exploits that fact to generate a large pool of inexpensive labor. In exchange, these students are rewarded with a US visa and a comparatively well-paying job. Neither party expects these graduates to stay in academia, so the ability to follow orders, work to a deadline, and solve highly-abstract problems provide a much more tangible benefit than inculcating the creativity, communication skills, and bureaucratic know-how that differentiates successful professors. In many ways, including these reasons, I think the PhD is starting to be viewed by some as a "super-Masters degree" that simply requires more work or takes longer, rather than as requiring a distinct, different skillset. Note that this is based only on my observations of a top-50 US graduate program; this might be totally different in a top-4 program, for example. However, one data point I found particularly persuasive is that in our program with ~300 PhD students, we have fewer than 30 (and probably fewer than 10) students with US citizenship or green cards.
- noodlenotes 8y ago
  I've observed this myself and it's creating a weird dynamic in data science. More often, companies are requiring PhDs for their entry level data science positions. (Because there's a surplus of people who want to enter the field and it's difficult to tell who will ultimately be successful. Possessing a PhD is a good predictor of success.) But there are very few American-born students going on to get PhDs who are interested in industry because, well, why would they waste another 5 years in school when they can get software engineering jobs with only a BS or MS? According to this study [1], 38% of all data scientists are non-US citizens and nearly half of early-career data scientist are non-US citizens. It's weird that we're pushing US citizens out of data science. [1] http://burtchworks.com/wp-content/uploads/2018/05/Burtch-Works-Study_DS-2018.pdf http://burtchworks.com/wp-content/uploads/2018/05/Burtch-Wor...
  - neuromantik8086 8y ago
    When you say "data science" does that encompass traditionally educated statisticians and business intelligence folks? If you only count the practitioners out there who self-identify as data scientists you get a pretty incomplete sample of what the field actually is.
kkylin 8y ago
Part of this difference is due to how departments are funded: math departments teach a lot of "service" courses for non-math majors. So there are enough TAships to fund most students who want them. Most other programs do not have enough TA positions to support all the students they have. Source: I'm on the math faculty at a large public research university.
- Fomite 8y ago
  Indeed. By way of comparison, my department has zero TA lines.
throwawayjava 8y ago
I have experience with both mathematics and CS departments, and I have exactly the opposite opinion. I think the Old School Mathematics model is absolute garbage and often explicitly and intentionally exploitative. 1. It's not like those math Ph.D.s are sitting around reading textbooks with their advisors. They are TA'ing service courses. A lot of TA'ing. In addition to TA'ing, they are prepping for a high-stakes exam. The exam will wash out the 50-70% of people who are needed as temp TA labor but whom the department has no intention of advising through to a Ph.D. After 2 or 3 years, you pass the exam. Yay! But wait. Reading textbooks with your advisor? No! you need to take a whole slew of courses. Why? Because course enrollment == money and easy/interesting teaching assignments. And good luck getting tons of research done in-between your TA'ing and courses... Now you're done with courses! Yay! After TA duties, you can spend 100% of your time on research (aka 50% of a 40 hour work week aka 80% of a grad student work week). Except you only have 2 years left :-( So now it's on to 1-3 postdocs so that you can start doing, after 5 years, what that CS phd student was doing when they "hit the ground running": build a research agenda that resonates with a large enough set of researchers that it translates into a research job. -- Also, the "one singular result" style of research is unique not just to mathematics, but to a particular subset of mathematics. It's just not a good fit for most research problems, even in Mathematics. I've seen tons of well-composed dissertations focused on a singular, pointless problem. The hodgepodge dissertations don't have a "beautiful singular result" feel, but almost invariably contribute at least one piece of useful, new knowledge. Finally, worth noting that the best and brightest don't leave research; they might leave academia, but typically not research. TL;DR: If you can find an advisor who isn't a slave driver (the majority of them at good places), the CS model is 1000% better than the Math model. You might be hired as a researcher for someone else's research agenda, but at least you don't spend the first 1/3rd of the phd as temp teaching labor stressing over high-stakes exams.
- impendia 8y ago
  > The exam will wash out the 50-70% of people who are needed as temp TA labor but whom the department has no intention of advising through to a Ph.D. At which department? Speaking as a math professor in an American research university -- this is unusual. Pass rates at most places are reasonably high, and professors hope that as many people as possible pass. It is true that grad students are expected to TA (if they want funding); usually grant funding is not available to support them. But generally the teaching burdens are relatively modest, around ~15 hours a week including grading and office hours.
KKKKkkkk1 8y ago
My experience does not agree with what you wrote in your last paragraph. Writing code as a computer scientist is equivalent to doing lab work as a biologist or to building instruments as an experimental physicist. You can't do research in CS if you don't get your hands dirty, unless your research in theoretical computer science. In exchange for this, many doors open to you after you finish your PhD that would otherwise be closed (unless you did your research in theoretical computer science). And among the best and the brightest that I met in industry, many are in fact PhDs.
setzer22 8y ago
As someone who's also pursuing a PhD in CS, I feel your comment doesn't reflect my situation, or my colleagues'. It's true that I started working on one of my advisor's projects, but the difference between the industry and academia is that I can discuss whatever I want about the project with him, and he will care about it. I am in lead of the project and my advisor acts as an advisor, not a boss. Of course it all depends on many factors, just my personal experience here.
joncrane 8y ago
I remember in undergrad I took a class and the topic of the day was the concept of an inflection point. The way the professor demonstrated it was like this: Based on time spent and potential earnings forgone, a Bachelor's degree is a good investment. A Master's is a good investment. The inflection point happens, and a Phd is not as good of an investment. That's all the information I needed to decide to stop after a Master's. Have only regretted it once, when I worked at a company that was owned by someone with a Phd in Economics and with many Phd Economists working there. I felt like a second class citizen and in a moment of irrationality, when the entire employment universe seemed to consist of only this company and ones like it, I thought about going and getting a Phd in my late 30s. Thank God I didn't. Making about 2.5x as much as I was making then, two jobs later, no further education needed (other than Cloud Guru courses).

gms7777 8y ago

I think different schools have different philosophies as far as that goes. There are schools that admit to the department and have students do very little research for their first two years, and there are schools that admit by advisor and you start research immediately. My PhD was the latter, but for me it was actually one of the major selling points for me on the program. I had a pretty good idea of the area that I wanted to work in, and I was always bad at learning from classes as opposed to hands-on self-driven learning. By the end of two years, I'd already gotten a lot of research done and moving towards the end of my PhD I was able to branch out within my area and take my own path in a way I doubt I would have been able to do in a different style program, where I wouldn't have had the same depth of knowledge in my field. All that to say, I think these are two different philosophies in designing a grad program, and each works better for a different style of student. To those that are still thinking about applying, its something that I think should be a serious consideration when looking at program fit.

throwawaymath 8y ago
With regard to your first paragraph, I don't think there are any American universities that admit students by advisor. As far as I'm aware, your admission is a decision made solely by the graduate admissions committee. Even if you have an advisor ready to take you on the most they can do is recommend you to the committee, not supplant it. When you're admitted there are still criteria aside from research to work on. I think this is a fundamental difference between American and European universities. In European universities, you can hypothetically join a PhD program without any prior degree at all if an advisor chooses to take you on personally. No such sidestepping can happen in an American institution, and regardless you'll have to take some courses (even if you're only auditing them in effect) before you can begin research. Unsurprisingly, American PhD programs are longer.
- overeater 8y ago
  Sorry to be blunt throwawaymath, but that's just wrong. I am very familiar with several American universities (including Ivy League) where the advisor can admit directly.
  - throwawaymath 8y ago
    I don't mind you being blunt at all, but do you mind naming those schools specifically? I'm still skeptical.
    - csa 8y ago
      Not OP, but Harvard, Yale, and Penn are three that I can speak to with firsthand experience. I will also say that, depending on the specifics of the applicant, you might be shocked by what admissions standards can be bent/broken based on the desire of a faculty member to have a specific student. It’s not common, but it happens.
- davmre 8y ago
  I've been on the PhD admissions committee for the Berkeley CS PhD program. Students in many areas are, for all practical purposes, admitted by an advisor. It's not a formal arrangement in the same sense as European programs, but no one gets admitted unless at least one professor will commit to being willing to advise and fund you. For superstar students this is rarely an issue and you have your pick of advisors, but for many admits there is exactly one professor who's identified your background and interests as a good fit, so you're essentially admitted with the expectation you'll work with that professor. Nothing you say is wrong; you're correct that everyone has to do course requirements. But culturally the course requirements are treated as a distraction from research, and students are expected to start working primarily on research from their first semester. I think this is more or less the case in all top American CS PhD programs, and there are reasons to set things up this way, but it's a very different environment from PhDs in most other fields.

chrisseaton 8y ago

What you're describing as what you didn't expect is how PhDs were done traditionally. It's the more recent US model to have classes and exams. I posed a question to my advisor in order to get onto the PhD, started researching immediately, and my 'general exam' half-way through was a presentation and viva. Obviously the question changed totally by the time I finished, but that's the nature of research and where it takes you. I'm not sure why you'd expect to explore computer science as a whole on a PhD. The point of a PhD is to become an expert in one part of the field and to push it to expand what we know - not to gain general wide knowledge. I don't think that was ever the idea - that's what a bachelors is for.

awelkie 8y ago
I think it depends on the field. In mathematics I think the expectation is that students need to study for a few years to even understand what the open questions are, and only then can they begin the research. I'm fully aware of what the point of a PhD is. I wasn't expecting to become a generalist, but I was expecting some time to study some recent results and come up with a tractable research proposal. Indeed, if the purpose of a PhD is to expand the totality of human knowledge, why was it expected that I would publish 3-4 conference papers to graduate? Wouldn't one significant result have been enough?
- cvwright 8y ago
  Often one "significant result" has to be serialized into 3-4 conference papers in order for it to be digestible and reviewable by the referees. There are exceptions of course, like Gentry's fully homomorphic encryption, but those are increasingly rare.
- gajjanag 8y ago
  > Indeed, if the purpose of a PhD is to expand the totality of human knowledge, why was it expected that I would publish 3-4 conference papers to graduate? This is a question that bothers me quite a bit with respect to EECS (electrical engineering and computer science). In mathematics, one's PhD thesis can often serve as one's first paper (once revised). Funnily enough, the same was true for EECS even at the top tier places in the 1980's - I know of a case at Caltech where the student did not have any publications prior to graduation, and yet went on to become a successful professor. Is it really true that humanity is producing more knowledge in EECS today per student than it was in the 1980's? I seriously doubt this. In my experience, instead what has happened is that the average "delta" in a conference proceeding has gone down significantly, in spite of the vast increase of words like "novel" and "new" being used, and the ballooning of the average paper length. In mathematics on the other hand, top quality authors freely admit that a lot of their work can be "implicitly" traced back to the "big names" of the past; indeed this has to be done for conceptual and historical clarity. I see that you mention Peter Sarnak in your top comment. Peter Sarnak has a a wonderful article at the very end of "The Princeton Companion to Mathematics" on his advice for students. Among other things, he stresses the importance of the "history of ideas", and how it often brings clarity to an entire field.

nbeleski 8y ago

Too many CS postgraduate programs just feel like engineering ones. This is why chosing a proper advisor is so important as your intentions should be aligned. Even then, there are things you might need to do to satisfy your program requirements that aren't exactly what you want to do. Anecdoctal experience: I was offered a somewhat decent scholarship for a Master's program backed by an Oil company, but I would work on specific projects chosen by the company. I sincerely prefered to go for a program/advisor that fit my goal (more on the theoretical CS research) and no grant (the program is free here). This meant that I have to work to pay my bills and study on 'free time'. This works for me personally because I make a more than any program scholarship would pay me while acquiring some industry experience. This can work while I am doing my masters, but It is borderline impossible for a PhD program.

pgbovine 8y ago

i totally know how you feel since i have felt that way in the past as well :) here's something relevant: http://pgbovine.net/practical-reason-to-pursue-PhD.htm http://pgbovine.net/practical-reason-to-pursue-PhD.htm my two cents: find something where the grants align with your interests, or find a brand-new faculty member who has startup money so isn't as bound by grants in the near term. and if things don't work out, simply leave and go to industry -- with a marketable skill set like what you get from being a CS major, there's no way for you to lose. go wherever has work that you like more.

mathattack 8y ago

I think the two subjects are very different. Math has a much longer history. To get to the edge of knowledge you need to know a lot more because people have built proofs upon proofs for much longer. You need an undergrad plus a year or two of grad to get both the depth a breadth of what’s been done in the past. In CS the edge of the field isn’t as far away. New knowledge invalidates more of the old, and you don’t even need a CS undergrad degree to be near some of the edges.

grigjd3 8y ago

It's pretty rare that PhD programs let you truly explore what you want to do. PIs (generally profs) get the vast majority of outside funding so you pretty much have to find a way to align yourself with a prof from day one.

chuckkir 8y ago

In general a professor is an entrepreneur. They are running a small business where they sell their product ("knowledge") to customers (business and industry) by means of sales and marketing (conferences and papers). Grad students are employees of this business, and of course the prof's incentive is to keep the good ones and fire (graduate) the bad ones, thus the reward for the student working hard is more hard work.

internetman55 8y ago
But no good student who approaches the grad school process in a sane manner is going to work for a professor whose students stay there for 12 years and don't graduate. A common thing for deciding if you want to work with someone is seeing how long their students took to graduate, where they ended up, emailing them, etc.
Fomite 8y ago
Two notes: 1) I am evaluated on the progress of my students. While bad ones will actually be fired, good ones are more useful to me if they go off and get postdocs etc. Some of the most powerful "businesses" in the field are created by the network of people who all trained with the same PI and then went on to found their own labs. 2) A burnt out, trapped graduate student is not actually one that produces good work, in my experience. 3) Several grants, which are how I fund my lab, will look very poorly on trainees making no progress. There's very little exploitation that's worth a program officer at the NSF wondering if I'm worth there time. Are there bad, exploitative PIs? Absolutely. But these people are assholes, not the only logical outcome of the system.

gervase 8y ago

> But also because it makes the PhD not much different than working in industry. I believe this is viewed as a feature, and not a bug. For the most part, academia serves as one of the few reliable ways to emigrate to the U.S., and exploits that fact to generate a large pool of inexpensive labor. In exchange, these students are rewarded with a US visa and a comparatively well-paying job. Neither party expects these graduates to stay in academia, so the ability to follow orders, work to a deadline, and solve highly-abstract problems provide a much more tangible benefit than inculcating the creativity, communication skills, and bureaucratic know-how that differentiates successful professors. In many ways, including these reasons, I think the PhD is starting to be viewed by some as a "super-Masters degree" that simply requires more work or takes longer, rather than as requiring a distinct, different skillset. Note that this is based only on my observations of a top-50 US graduate program; this might be totally different in a top-4 program, for example. However, one data point I found particularly persuasive is that in our program with ~300 PhD students, we have fewer than 30 (and probably fewer than 10) students with US citizenship or green cards.

noodlenotes 8y ago
I've observed this myself and it's creating a weird dynamic in data science. More often, companies are requiring PhDs for their entry level data science positions. (Because there's a surplus of people who want to enter the field and it's difficult to tell who will ultimately be successful. Possessing a PhD is a good predictor of success.) But there are very few American-born students going on to get PhDs who are interested in industry because, well, why would they waste another 5 years in school when they can get software engineering jobs with only a BS or MS? According to this study [1], 38% of all data scientists are non-US citizens and nearly half of early-career data scientist are non-US citizens. It's weird that we're pushing US citizens out of data science. [1] http://burtchworks.com/wp-content/uploads/2018/05/Burtch-Works-Study_DS-2018.pdf http://burtchworks.com/wp-content/uploads/2018/05/Burtch-Wor...
- neuromantik8086 8y ago
  When you say "data science" does that encompass traditionally educated statisticians and business intelligence folks? If you only count the practitioners out there who self-identify as data scientists you get a pretty incomplete sample of what the field actually is.

kkylin 8y ago

Part of this difference is due to how departments are funded: math departments teach a lot of "service" courses for non-math majors. So there are enough TAships to fund most students who want them. Most other programs do not have enough TA positions to support all the students they have. Source: I'm on the math faculty at a large public research university.

Fomite 8y ago
Indeed. By way of comparison, my department has zero TA lines.

throwawayjava 8y ago

I have experience with both mathematics and CS departments, and I have exactly the opposite opinion. I think the Old School Mathematics model is absolute garbage and often explicitly and intentionally exploitative. 1. It's not like those math Ph.D.s are sitting around reading textbooks with their advisors. They are TA'ing service courses. A lot of TA'ing. In addition to TA'ing, they are prepping for a high-stakes exam. The exam will wash out the 50-70% of people who are needed as temp TA labor but whom the department has no intention of advising through to a Ph.D. After 2 or 3 years, you pass the exam. Yay! But wait. Reading textbooks with your advisor? No! you need to take a whole slew of courses. Why? Because course enrollment == money and easy/interesting teaching assignments. And good luck getting tons of research done in-between your TA'ing and courses... Now you're done with courses! Yay! After TA duties, you can spend 100% of your time on research (aka 50% of a 40 hour work week aka 80% of a grad student work week). Except you only have 2 years left :-( So now it's on to 1-3 postdocs so that you can start doing, after 5 years, what that CS phd student was doing when they "hit the ground running": build a research agenda that resonates with a large enough set of researchers that it translates into a research job. -- Also, the "one singular result" style of research is unique not just to mathematics, but to a particular subset of mathematics. It's just not a good fit for most research problems, even in Mathematics. I've seen tons of well-composed dissertations focused on a singular, pointless problem. The hodgepodge dissertations don't have a "beautiful singular result" feel, but almost invariably contribute at least one piece of useful, new knowledge. Finally, worth noting that the best and brightest don't leave research; they might leave academia, but typically not research. TL;DR: If you can find an advisor who isn't a slave driver (the majority of them at good places), the CS model is 1000% better than the Math model. You might be hired as a researcher for someone else's research agenda, but at least you don't spend the first 1/3rd of the phd as temp teaching labor stressing over high-stakes exams.

impendia 8y ago
> The exam will wash out the 50-70% of people who are needed as temp TA labor but whom the department has no intention of advising through to a Ph.D. At which department? Speaking as a math professor in an American research university -- this is unusual. Pass rates at most places are reasonably high, and professors hope that as many people as possible pass. It is true that grad students are expected to TA (if they want funding); usually grant funding is not available to support them. But generally the teaching burdens are relatively modest, around ~15 hours a week including grading and office hours.

KKKKkkkk1 8y ago

My experience does not agree with what you wrote in your last paragraph. Writing code as a computer scientist is equivalent to doing lab work as a biologist or to building instruments as an experimental physicist. You can't do research in CS if you don't get your hands dirty, unless your research in theoretical computer science. In exchange for this, many doors open to you after you finish your PhD that would otherwise be closed (unless you did your research in theoretical computer science). And among the best and the brightest that I met in industry, many are in fact PhDs.

setzer22 8y ago

As someone who's also pursuing a PhD in CS, I feel your comment doesn't reflect my situation, or my colleagues'. It's true that I started working on one of my advisor's projects, but the difference between the industry and academia is that I can discuss whatever I want about the project with him, and he will care about it. I am in lead of the project and my advisor acts as an advisor, not a boss. Of course it all depends on many factors, just my personal experience here.

joncrane 8y ago

I remember in undergrad I took a class and the topic of the day was the concept of an inflection point. The way the professor demonstrated it was like this: Based on time spent and potential earnings forgone, a Bachelor's degree is a good investment. A Master's is a good investment. The inflection point happens, and a Phd is not as good of an investment. That's all the information I needed to decide to stop after a Master's. Have only regretted it once, when I worked at a company that was owned by someone with a Phd in Economics and with many Phd Economists working there. I felt like a second class citizen and in a moment of irrationality, when the entire employment universe seemed to consist of only this company and ones like it, I thought about going and getting a Phd in my late 30s. Thank God I didn't. Making about 2.5x as much as I was making then, two jobs later, no further education needed (other than Cloud Guru courses).