Master's and Ph.D. degrees

[Nick]

Hi all,

I'm curious about the difference between having a Masters or a Ph.D. degree in this industry.

Does a Ph.D. automatically get valued higher, or does experience and personality determine everything? Does a Ph.D. only end up in management or do they 'get their hands dirty' as well? Is is near impossible to reach the top of the ladder with a Master's degree? Is a Ph.D. only interesting for academic research, or does it bring a significant advantage in knowledge for current development? Is it of great importance where the Ph.D. degree was obtained?

Sorry for the many questions. I will probably recieve my Master's degree this summer, and my thesis promoter would like me to continue my research. While I like the idea of working on my own projects for several years, I don't want to get too detached from the industry either. Basically, I don't know whether a Ph.D. degree would broaden my possibilities, or narrow them. Or does that depend on other factors?

Thanks for the opinions and advice,

Nicolas

 

[stevem]

Tough decision. A PhD is notionally training for academia. It will depend on individual employers as to whether they value a PhD. In my area(s), MBAs are valued more highly by employers than other graduate degrees. Anecdotally, I had previously tutored a number of employees with MBAs at a firm I once worked for. All were hired at a higher salary level than me...

 

[Nick]

Tough decision. A PhD is notionally training for academia. It will depend on individual employers as to whether they value a PhD. In my area(s), MBAs are valued more highly by employers than other graduate degrees. Anecdotally, I had previously tutored a number of employees with MBAs at a firm I once worked for. All were hired at a higher salary level than me...

Thanks for the information. I'm not interested in an MBA at all, nor does salary influence my descision to continue my academic education. I just want to be able to work on cutting-egde technology and do what I think is best (yeah I'm a rebel). My main interest is graphics and compiler technology, and a bit of digital design.

The chief engineers, the true 'inventors' at companies like AMD, Intel, NVIDIA, what kind of degrees do they have, and how relevant is that? Ok, maybe I'm aiming too high, but I just don't want to regret my descision later...

 

[MrGaribaldi]

Going by what one of my prof's told me about his experiences getting hired (in the games industry) with a Ph.D. it can be more of a hindrance than a help.
The reason for this is, as stevem said, that Ph.D.'s are considered more of academics than actual implementors. They are thought to sit in their Ivory towers and research graphics, but either aren't very interested in real world applications of said research or just don't know how to deal with practicle applications (which are true for some, but not for others)
So you might have to fight harder to get your foot in the door, though, as more and more Ph.D's make their way into the wild, that should get easier.

And from what I gathered, when you've gotten the foot in the door, it can give a boost when climbing the corporate ladder.

[EDIT]
Didn't think/realize it was the "pure" graphics industry you were interested in,
so this post isn't really relevant...
[/EDIT]

 

[armchair_architect]

I think it depends a lot on individual corporate cultures and the kind of job you want.

If the company is looking for a code monkey, or even someone doing design work in a fairly well-understood area (yet another e-commerce system, for example), a Ph.D. will probably be a hindrance.

In my experience, in companies focused more on advancing the technological state of the art, a Ph.D may or may not help you but won't hurt compared to an M.S. The important thing is to demonstrate that you're not an ivory-tower academic. As part of your thesis, or in your free time (hah!), *build real systems*. Show that you can start and finish a project that involves real engineering tradeoffs. If you can do that, you'll get hired. Companies like this value creative, innovative people, but need their ideas to be practical, and ideally the inventor can be heavily involved in the conversion from idea to product.

Of course, if you can show you're that kind of person they'll welcome you even without the Ph.D.

 

[silent_guy]

I just want to be able to work on cutting-egde technology and do what I think is best (yeah I'm a rebel). My main interest is graphics and compiler technology, and a bit of digital design.

The chief engineers, the true 'inventors' at companies like AMD, Intel, NVIDIA, what kind of degrees do they have, and how relevant is that? Ok, maybe I'm aiming too high, but I just don't want to regret my descision later...

It's incredibly dependent on the company and on the person in question. There are a lot of companies for whom a PhD is not a reason to (initially) pay more, because there's little proof that an average PhD is smarter at getting the things done that need to be done: a lot of work in tech requires a healthy understanding and intelligence, but it doesn't require a herd of geniuses to do the grunt work. It's different for truly innovative companies that are really at the cutting edge: some of the big ones, but also (especially!) startups with a revolutionary idea that can be a main competitive advantage. Speaking about compiler technology, a good example would be a compiler that is specifically targeted towards parallel architecture. A PhD in that field could be worth a lot.

Second part is the person itself: it's never a bad idea to know how to get your hands really dirty as a PhD. Unless you're the one-in-a-million genius with a bag of landmark research papers who is so obviously brilliant that using him for grunt work would be an incredible waste, you're going to be deployed to work on fun stuff as well as just the boring stuff to get the product out. (Insert Thomas Edison quote about inspiration and perspiration here.) So make sure you present yourself as somebody with a lot of practical experience, not just theoretical knowledge.

If you're good enough, interesting work will present itself once you've established a reputation, Phd or no Phd.

Finally, it's never a bad idea to be able to have something to show when you first enter the job market: interesting stuff you've done. Programming prototypes. Open source stuff. It's an very good way to stand out of the crowd.

Edit: One more thing: make yourself visible on the web under your real name. I don't always google somebody when I get a resume, but if the resume hints at it, I definitely will. Don't bother, of course, if I'll see 100 links with nonsense and flame fests, but 100 links with useful advice in, say, a compiler email list or a highly technical newsgroup or, say, a bunch of interesting book reviews on Amazon, it will help.

(Edit: @armchair_architect: looks like we have exactly the same opinion. )

 

[Nick]

If the company is looking for a code monkey, or even someone doing design work in a fairly well-understood area (yet another e-commerce system, for example), a Ph.D. will probably be a hindrance.

The last thing I want to be is a code monkey. Don't get me wrong, I love coding, and sometimes you need to do boring things before you can get to the interesting stuff, but I don't want to do what thousands have done before me.

In my experience, in companies focused more on advancing the technological state of the art, a Ph.D may or may not help you but won't hurt compared to an M.S.

Interesting. So is it ever a hindrance not to have a Ph.D. degree? I understand that with an M.S. you still have to prove yourself in certain areas, but are there situations / companies where you'd never get the same the same opportunities as someone with a Ph.D. degree?

As part of your thesis, or in your free time (hah!), *build real systems*. Show that you can start and finish a project that involves real engineering tradeoffs.

My Master thesis was to extend SoftWire with a high-level front-end and an optimizing middle-end, so I now call it SoftWire++. The commercial branch of SoftWire is used in SwiftShader, of which I'm the lead developer.

Of course, if you can show you're that kind of person they'll welcome you even without the Ph.D.

They'd welcome me, but would I get the same opportunities?

Thanks for your opinion!

 

[Nick]

Speaking about compiler technology, a good example would be a compiler that is specifically targeted towards parallel architecture. A PhD in that field could be worth a lot.

Actually that's not so far from what I'd like to research. So do you think it would be a missed opportunity not to do it?

 

[silent_guy]

Actually that's not so far from what I'd like to research. So do you think it would be a missed opportunity not to do it?

Just today I saw a front-page article on EETimes deploring the lack of computer scientists researching parallelization in general. But whatever you decide to do, there'll be a million other missed opportunities.

The last thing I want to be is a code monkey. Don't get me wrong, I love coding, and sometimes you need to do boring things before you can get to the interesting stuff, but I don't want to do what thousands have done before me.

No matter what bright idea you're going to come up with, somebody will have to implement it and there's a 99% chance that it's going to be you. Most likely the final implementation, but certainly the proof of concept and since most easy problems are already solved, you're going to be coding (and debugging!) at lot anyway. "Code Monkey" is a very vague term.

Interesting. So is it ever a hindrance not to have a Ph.D. degree? I understand that with an M.S. you still have to prove yourself in certain areas, but are there situations / companies where you'd never get the same opportunities as someone with a Ph.D. degree?

It's definitely a hindrance if you want to do something in the academic world. It can open certain doors more quickly if your PhD subject is in exactly the right field. Some companies want the best and the brightest: while recruiting among a PhD population doesn't guarantee anything, it increases their chances. Google comes to mind...
But a lot of tech projects requires large teams working together for a long time. In those cases, prima donna's aren't nearly as useful as reliable group workers but also there the good guys are known very quickly (and get the most challenging work.)

My Master thesis was to extend SoftWire with a high-level front-end and an optimizing middle-end, so I now call it SoftWire++. The commercial branch of SoftWire is used in SwiftShader, of which I'm the lead developer.

Well, you can at least put a major check mark behind 'relevant presence on the web' then.

They'd welcome me, but would I get the same opportunities?

I believe it won't make a major difference. It is surprising how quickly reputations are established in a company. If a company fails to act accordingly, you're better off going somewhere else.

 

[stevem]

Thanks for the information. I'm not interested in an MBA at all, nor does salary influence my descision to continue my academic education. I just want to be able to work on cutting-egde technology and do what I think is best (yeah I'm a rebel). My main interest is graphics and compiler technology, and a bit of digital design.

I believe a PhD degree should be pure research, beholdent to innovation only - otherwise a vocational training degree will suffice... I was just pointing out the sunk cost of innovation is sometimes a subordinate position in a group, not the leadership position many have expectations of. In my case, I took the job because I was interested in the project that was starting up. As the project was wound down, most of my colleagues were let go.

The chief engineers, the true 'inventors' at companies like AMD, Intel, NVIDIA, what kind of degrees do they have, and how relevant is that? Ok, maybe I'm aiming too high, but I just don't want to regret my descision later...

You won't. You'll only kick yourself for not taking the opportunity if presented. I see it time & time again. I'm currently considering co-supervising the PhD project of a colleague who never got around to submitting his original thesis ~35yrs ago. His motivation is certainly not an increase in salary or position, he's a senior partner & director...

The one piece of advice I can offer is "strike while the iron is hot". Most topics/areas of expertise have temporal elastic demand 'til one day they're de rigueur. If you can differentiate yourself from the crowd by taking advantage of ~2yrs sunk cost of Masters, then I'd say do it. Keep track of the outside world, though. A simple e-mail to outside groups can work wonders. Ask if you can go for an informal visit, talk about common areas of interest, etc. Obvious stuff.

Good luck.

 

[armchair_architect]

So is it ever a hindrance not to have a Ph.D. degree? I understand that with an M.S. you still have to prove yourself in certain areas, but are there situations / companies where you'd never get the same the same opportunities as someone with a Ph.D. degree?

At an established company, working on something the company is good at, I don't think so. If you demonstrate you're good at what they do, they'll be happy to give you the opportunity to help them be more competitive! The places I've worked at that build leading-edge products are very much meritocracies -- if you're talented, you'll advance and be asked to work on the high-importance projects (which are usually the most interesting), regardless of what initials you have behind your name. The company would be stupid to work any other way.

I can't speak from experience here, but the places where not having a Ph.D might hurt are corporate research organizations (Microsoft Research, IBM's T.J. Watson research group, etc.), and startups. The former is very similar to academia. The latter will typically be looking for a mix of experienced developers who have a track record of getting things done on small teams, and recognized domain experts (i.e. Ph.Ds with highly relevant thesis/publications). This also goes for established companies trying to get into new areas where they don't already have a lot of expertise, since those teams often have a somewhat startup-like structure.

 

[Gelanin]

I dont know much about this at all, but i can well imagine that research into parallellism and/or a compiler that made it easier to get multithreaded code/applications would become extremely useful in the future considering how it all seems to point towards more cpu/cores instead of faster ones.

 

[Nite_Hawk]

Hi all,

I'm curious about the difference between having a Masters or a Ph.D. degree in this industry.

Does a Ph.D. automatically get valued higher, or does experience and personality determine everything? Does a Ph.D. only end up in management or do they 'get their hands dirty' as well? Is is near impossible to reach the top of the ladder with a Master's degree? Is a Ph.D. only interesting for academic research, or does it bring a significant advantage in knowledge for current development? Is it of great importance where the Ph.D. degree was obtained?

Sorry for the many questions. I will probably recieve my Master's degree this summer, and my thesis promoter would like me to continue my research. While I like the idea of working on my own projects for several years, I don't want to get too detached from the industry either. Basically, I don't know whether a Ph.D. degree would broaden my possibilities, or narrow them. Or does that depend on other factors?

Thanks for the opinions and advice,

Nicolas

Hi Nick,

I only have my bachelor's degree, so take my opinion with a degree of salt. I used to work in government contracting, and now I work for the Minnesota Supercomputing Institute. Here's what I've seen. In government contracting, having your Ph.D got you more money and a better chance at interesting work, but it also means that you are the first person to be let go when times get tough. Why? Because when funding for your project dries up and you need new work, they would rather keep someone cheap who is happy doing boring work than someone expensive who hates it.

At the University things are different. There is a much greater divide between the employees here with Ph.Ds and those without. They would never lay off someone here with a Ph.D and replace them with someone with a bachelor's degree. That doesn't mean that you need a Ph.D here to do interesting work, but you have to be smarter to get it. People with Ph.Ds are assumed to have already paid their dues, while those of us without have to prove our intelligence.

If I were hiring you, the Ph.D would only carry as much weight as the quality of your work. I know too many people who have breezed through school without learning much of anything. On the other hand, if you did good work the Ph.D could be a big selling point. The biggest advantage the Ph.D gives you in my humble opinion is the (presumably) successful execution of a major project. Being able to independently define and solve a difficult problem is a huge selling point.

In the end I really wouldn't worry about it too much. If you like doing research and you like what you've been doing with your master's degree then keep going and get your Ph.D. If you are tired of academia start applying around and see what you can get. If nothing turns up, stay in academia and try to build up some industry contacts while continuing your research. Definitely start talking to people now though.

Good luck!
Mark

 

[ban25]

They'd welcome me, but would I get the same opportunities?

You should look at it more from the perspective of what you want to do. PhD's are generally for academia, as has already been mentioned here. If you want to become a Professor at your University or another, you will almost certainly need it (and even if not, you'll get better pay). This also holds true for research organizations in the industry. But otherwise, it's not really going to make any difference. Without substantial team lead experience, you won't be offered a management position and without substantial work experience, it may be hard to get a senior or architect-level position. It might even be harder to get entry and mid-level positions if for no other reason than salary expectations. So you'll have to climb the ladder like everyone else, but you may have a modestly higher starting-salary (depends on the company, but in any case, it won't offset the cost of the degree).

The catch is, if you like pure research, academia is one of the only (and best) places to do it. So if that's your thing, then the PhD is definitely worthwhile. And if you do some high-profile work while you're there, then that will go a long way towards getting your name out there.

 

[Humus]

While I like the idea of working on my own projects for several years, I don't want to get too detached from the industry either. Basically, I don't know whether a Ph.D. degree would broaden my possibilities, or narrow them.

What makes you happy, climbing the corporate ladder and earn heaps of money or working on the forefront of technology? If the former, don't waste your time doing research, a master's degree is more than plenty if you're a manager type person. The average CEO may be more technical than average, but is probably not someone who'd know how to write a shader for instance. If you enjoy research and development on the other hand, then stay in the academic world as long as you can. It's hard to find a position in the corporate world where you'll enjoy the same amount of freedom and openness.

 

[Nick]

What makes you happy, climbing the corporate ladder and earn heaps of money or working on the forefront of technology?

Money doesn't drive me, but I love working with cutting-edge technology as long as it's practical. In other words, I'm interested in science for the technology, not science for science.

If you enjoy research and development on the other hand, then stay in the academic world as long as you can. It's hard to find a position in the corporate world where you'll enjoy the same amount of freedom and openness.

Sounds like I could fit in there.

I'm still somewhat undecided though. There are lots of opportunities for me in the industry right now. A doctorate could mean I'm missing my chance to work on interesting new things. There are really exciting changes taking place in CPU and GPU development right now... so shouldn't I strike the iron while it's hot? Let me put it this way; I'm a little more interested in what the industry produces than what academia produce, although I do understand the necessity of the latter.

Thanks a lot for the opinions!

 

[Graham]

Nothing is stopping you from doing research in an non-academic setting, or a mixed academic/commercial setting. It could even be a benefit in some ways - you may actually have more control over the technology and the research choices. You also don't have to commit to the PhD, which may well be a 3 year idea that you discover is not feasible - or already solved - after 6 months.

I'm in a bit of an odd position. I got my university degree, however I didn't exactly give it my full attention. During the time, I was also working on projects for a research lab within the university, and many of my own projects. This got me noticed by visiting academics, which got me into an even bigger project. It ended up meaning I got my current job (that I start in a week) which will involve easily the most ridiculous (yet awesome) project so far. This is for a more traditional game company.. That said I've also had job offers from puerto rico, an aquarium in singapore, to US military R&D. I don't think I'd be where I am had I continued my studies - however I'm kinda weird.

A similar thing happened to a friend at the lab, he was year or two ahead of me - so he has an honours degree - He is still there, but to my knowledge he will be moving to the US to work for a large sci-fi board game company, doing 3D research work (trend spotting, so to speak). He met these people while attending siggraph. He was attending siggraph because of work he'd done on a an project unrelated to his degree.

Finally, one guy, who I know quite well, decided to take on a PHD. 2+ years in, he's depressed and frustrated with it. The problem is that it simply isn't fresh anymore. In fact, come to think of it, everyone I know who are working on PhDs are frustrated with them...

The problem with the technology sector is that research needs to be very rapid and iterative. PhDs are neither (imo). Consider what was cutting edge 3 years ago, and it's not so pretty.

I think you should at least have a look into your options for paid, or internship research on actual commercial projects (6 month duration or less). Overseas work especially. Travelling about and getting your name out there (by producing cool things) can only be good for your future project prospects. Unfortunately these kinds of institute are rare, and often times they are more interested in self preservation or indulgence than actually advancing knowledge.

If I were hiring you, the Ph.D would only carry as much weight as the quality of your work. I know too many people who have breezed through school without learning much of anything. On the other hand, if you did good work the Ph.D could be a big selling point. The biggest advantage the Ph.D gives you in my humble opinion is the (presumably) successful execution of a major project. Being able to independently define and solve a difficult problem is a huge selling point.

This man speaks truth profound. However it is nothing you cannot do outside of a university setting (you just don't get the three extra letters to your name)

 

[Techno+]

PHd isnt important even if u r becoming a top executive at companies like AMD, Intel etc. Phil Hester, the CTO of AMD, has a bachelors and masters degree in EE, but still he is top CTO becuase of his incredible work when he was at IBM.

 

[retrox]

But whatever you decide to do, there'll be a million other missed opportunities.

The truth is there. Don't regret anything, or you'll regret everything. Go into whatever interests you most, and make great things. Some time down the road, what you find really exciting now won't be that interesting or relevant anymore. It won't matter. What will matter is what you achieved.

I remember discussions on opengl.org forums several years ago. The exciting thing back then was register combiners. There were highly regarded contributors who were discussing problems and how to solve them with combiners. All this stuff is irrelevant now. But these people certainly had a great time working on these issues and most likely gained recognition in whatever project / organization / company they were at that time when they came up with their solutions.