D3veate

This is a question for those of you out there who prefer computer science over software engineering: was there one thing that made you prefer academia over industry?
 
This is what made me favor academic pursuits. For the last semester of my master's program, I spent 40 hours a week as a software engineer and spent my remaining time working on my thesis. That was the experience that made me realize that I preferred research. It was difficult to walk home after a full day of staring at a computer screen just to resume staring at a computer screen for another five or so hours. However, throughout all of this, I wasn't able to care much about my work projects. In contrast, research questions kept popping into my head and consuming all of my attention.
 
Ultimately, my preference for academia comes down to one issue. Academic projects don't need a business justification; I prefer to work on the most interesting project.
 
 
 
What experience made you prefer academia?

Here is something that is similar to what you're asking for: http://cs.brown.edu/people/alexpap/faculty_dataset.html

One of the links there has a spreadsheet of 2195 professors. Associated with each name is the professor's current
institution and his or her subfield. Furthermore, each name has a link to a website where you should be able to
obtain more information about that professor's research.

I've always assumed that getting into a PhD program is about convincing some professor to work with you on research. GPA and GRE numbers work decently as filters. You can convince a professor that you're capable of quality research because you have already demonstrated a proclivity for doing good research, or because you bring a valuable skill to a research group, or because your professors can vouch that you have what it takes to succeed in academia, or because your enthusiasm and goals indicate that you are a researcher, albeit one who needs training.

The numbers you've posted suggest that you'll get past most of the filters. However, since you've already visited Duke and don't already know that you have a place in a lab, I assume that you didn't have a chance to talk with (and hit it off with) a prospective adviser. Duke might be where you pursue a PhD, but it isn't the only place doing robotics and AI research. It isn't the only place where your fellow graduate students will have the intelligence, skills, and academic curiosity that will enrich your own experience.

Your chances of getting in are either 0 or 1; you'll know after you apply which one it was. A lot of qualified people will apply. The probability of an arbitrary qualified applicant being accepted is Pr(accepted | qualified) = (#accepted) / (#qualified). This of course assumes that Pr(qualified | accepted) = 1. If you apply to a lot of schools with similar admission standards, then we can estimate the probability that you're qualified for those schools... I can't think of any reason why that would be useful. Oh well.

Here's a question: does it matter what your chances are? If you want a PhD, you'll do yourself a disservice if you don't pursue it.

I've always assumed that getting into a PhD program is about convincing some professor to work with you on research. GPA and GRE numbers work decently as filters. You can convince a professor that you're capable of quality research because you have already demonstrated a proclivity for doing good research, or because you bring a valuable skill to a research group, or because your professors can vouch that you have what it takes to succeed in academia, or because your enthusiasm and goals indicate that you are a researcher, albeit one who needs training.

The numbers you've posted suggest that you'll get past most of the filters. However, since you've already visited Duke and don't already know that you have a place in a lab, I assume that you didn't have a chance to talk with (and hit it off with) a prospective adviser. Duke might be where you pursue a PhD, but it isn't the only place doing robotics and AI research. It isn't the only place where your fellow graduate students will have the intelligence, skills, and academic curiosity that will enrich your own experience.

Your chances of getting in are either 0 or 1; you'll know after you apply which one it was. A lot of qualified people will apply. The probability of an arbitrary qualified applicant being accepted is Pr(accepted | qualified) = (#accepted) / (#qualified). This of course assumes that Pr(qualified | accepted) = 1. If you apply to a lot of schools with similar admission standards, then we can estimate the probability that you're qualified for those schools... I can't think of any reason why that would be useful. Oh well.

Here's a question: does it matter what your chances are? If you want a PhD, you'll do yourself a disservice if you don't pursue it.

Is the prestige of the university important...

I wish there were a thread dedicated to this exact question.

Here are some observations I've made about this question. First, people in software engineering seem to be acutely aware of where everyone else went to school. I come from Hodunk U, and I've always had the impression that when people find out that they use this to judge what I'm capable of. I've had the liberty of talking to another software engineer who graduated from my university... if you were to guess where he went to school based on the projects he's produced, you would guess that his alma mater is something on the caliber of Stanford. He has made the same observation. Apparently it really irks people from elite universities when they realize that the work that he produces is better than the work that they produce.
 
In short: the prestige of the university does matter. It doesn't mean you'll be a better engineer, but will affect people's immediate perception of you. Of course, my experience comes from an extreme; it's quite likely that the SUNY system is prestigious enough that people will be inclined to wait to form their opinions of you until they've seen what you can do.

Here's the thing about computer science... In school, you learn how to be a scientist in an academic setting. Employers are looking for engineers who can get work done in an industrial setting. It's common to hear stories about newly hired PhD's who don't know how to program. I've known several software engineers who earned degrees in something other than CS. Computer engineering is a common source, as are mathematics and electrical engineering. I've met successful software engineers who earned their degrees in criminal justice or art.
 
The majority of the skills needed to be a successful software engineer are things learned outside of a classroom. My personal experience has been that only 20% of what I use in software engineering is a result of something I learned in a classroom. Don't get me wrong -- this is an important 20%, and you won't be able to be a top-notch engineer without a reasonable theoretical understanding of what you're working on.

To answer some of your specific questions... An MS is better than a BS. If an MS takes longer, you might want to get a second BS... or not even get a second degree. An MS in CS will serve as an admission voucher for the software engineering industry... but a proficiency with code and an art degree can also be sufficient. If you can demonstrate that you're a good engineer, you'll get good job offers. An MS will not prevent you from applying to entry level positions. An MS won't give you much of a head-start over someone with just a BS, however.

Data structures is certainly something you'll want to brush up on. There is an excellent chance that an technical interview you do will focus on two things:
 
Do you know how do use your language of choice? Can you reason through basic data structure problems? Another key difference is that some employers want to know your grade in compilers; this is a subject that isn't directly used by a lot of software engineers, but it's a universally difficult class, and the grade in that class gives a decent idea of whether you can handle large projects. This is particularly important because most class assignments are short, write-and-forget type programs. If you do a masters program, you should be able to point at something equivalent. If you can put a decently sized program on github, then employers won't need to look at your compilers grade to get an idea of whether you can handle larger programming projects. Of course, you can always take compilers as a graduate student.
 
The wonderful thing about software engineering is that everyone has holes in his or her knowledge. One of the items on the ACM code of ethics gives some guidance on this; a professional software engineer should always strive to acquire and maintain professional competence. Identify your weak areas and make a constant effort to improve yourself.