Can I Salvage A Poor Undergrad GPA?

[DorkRawk]

So I'm coming to the end of a fruitless PhD acceptance season (I'm 0 for 8 right now, waiting for one more school) and thinking about applying to programs next year. My main problem is that my undergrad GPA is terrible. Not bad. Terrible. Here's my background:

Undergrad in Math/CS at a top 5 CS school w/ a GPA around 2.3

Masters in Human Computer Interaction at smaller, lesser known school w/ a GPA of 3.98

My GRE scores are fine

1 published paper (not first author) in a small HCI conference

I'm interested in AI research, so I need to show that I can do strict CS work.

How can I make up for my undergrad GPA? Would taking the subject GRE help? How can I get more relevant research experience outside of a college environment? How can I figure out what schools would be realistic for me when I have a sort of non-standard background?

[fangsout]

I can only speak from an American perspective but it may be similar in other countries. You absolutely can make up for a poor undergrad gpa. In no particular order:

- Take relevant CS and Math classes as a non-degree seeking student at the nearest college or a certificate program over distance learning at a decent school.

- look for internships and apply to 100's of positions for jobs/projects that would give you valuable experience and take, if you can, the best opportunity you get regardless of location.

- seriously study for and take the CS GRE.

- contact professors, after thoroughly reading their papers, that are working on projects you find interesting and ask questions to try to figure out what specific topic or problem area you may be interested in.

[adinutzyc]

Also, try to see if you can gain some research experience... I don't really know how you could work this out, but publishing a paper, or presenting a poster at a conference would do wonders. You could start by contacting some professors on the NSF REU list (unless you have a job) and ask if they would need some help for the summer, explaining your circumstances and the reasons you would like to work with them. I can't guarantee they will say yes, but if they do and you do a good job, they would be able to give you a very strong recommendation. Although I was an undergraduate in a US institution, I had to do the same thing (I am an international student and we aren't really eligible for REUs), and there were more than a few people who were impressed by my willingness to devote my summer to doing research with them (I did it for free btw, because they only had the NSF money that I wasn't eligible for, so you have to have some money to invest in this)

[DorkRawk]

Thanks for the advice everybody. While I'm working on some research right now, I'm trying to decide how I want to make up for my poor CS undgrad grades. I'd prefer to take classes over taking the CS subject GRE. If I were to take classes, what classes should I take? (again, my undergrad grades were REALLY shitty and I think that admissions committees are worried about my ability to do grad level CS work) I've taken intro to AI, intelligent information retrieval, and machine learning/neural networks classes at the grad level as electives during my HCI masters programs so I have some CS grad experience. I'll likely be moving to Berkeley, CA soon, so I was hoping to take some classes as a non-degree seeking student there.

If I had to pick only a couple classes to take between now and the next application season, what should I take to best make up for poor undergrad grades to show committees that I have the foundation to do work in their program?

[Azazel]

I have a friend who got into a few Top 10 schools (including one of the Top 4) with around a 2.0 GPA.

How did he do it? Basically, he had a professor who aggressively went to bat for him. More than just writing an great letter, this professor made a bunch of phone calls and convinced his colleagues how awesome my friend would be as a graduate student.

So my advice to you would be to find a respected professor (ideally, one with connections at the schools to which you want to gain entrance), and convince that professor you would make an excellent researcher/academic. Usually, this will involve doing high-quality research under said professor ;-)

Good luck -- while hard, it's definitely possible to work around a low GPA.

[victor.s.andrei]

Thanks for the advice everybody. While I'm working on some research right now, I'm trying to decide how I want to make up for my poor CS undgrad grades. I'd prefer to take classes over taking the CS subject GRE. If I were to take classes, what classes should I take? (again, my undergrad grades were REALLY shitty and I think that admissions committees are worried about my ability to do grad level CS work) I've taken intro to AI, intelligent information retrieval, and machine learning/neural networks classes at the grad level as electives during my HCI masters programs so I have some CS grad experience. I'll likely be moving to Berkeley, CA soon, so I was hoping to take some classes as a non-degree seeking student there.

If I had to pick only a couple classes to take between now and the next application season, what should I take to best make up for poor undergrad grades to show committees that I have the foundation to do work in their program?

Several things you didn't mention: what type of graduate programs you were applying to (master's vs. PhD) and where (i.e., what schools).

The admissions committee's focus is going to differ based on the individual department and the type of degree.

If you were applying for a PhD (since you have a master's degree), then what might have bothered the admissions committees was your ability to conduct research, given your low GPA. What classes did you mess up in...algorithms? Data structures? Computer architecture?

You also didn't mention whether you did a master's thesis.

Also, Azazel is right - LORs make a big difference. One of my colleagues had a 2.8 but several REUs, including one where he impressed a professor who took a bet on him when others wanted to say no. He just got into one of the schools where did an REU, CS PhD with a full fellowship, no restrictions - and why? Because the professor he impressed went to bat for him.

[starmaker]

If I had to pick only a couple classes to take between now and the next application season, what should I take to best make up for poor undergrad grades to show committees that I have the foundation to do work in their program?

"Core" classes. Algorithms would be an excellent one, or possibly theory of computation. A computer/operating systems class would also be good.