Undergrad CS Research Question

[vnd1242]

Hello, 

I'm currently an undergrad math major considering a PhD in theoretical CS. I'm at a T15 but very industry oriented university, where undergrad research is more rare than at other similar schools. 

I have had two research experiences so far in theoretical CS, both of which consisted of me emailing 5-10 professors, them sending me a problem and 2-3 papers, and me struggling to solve the problem on my own. It will be difficult for me to solve them, but If I do end up solving one of these problems I'm guaranteed to be first author on a publication.

However, a lot of my friends considering grad school that are CS majors work for large laboratories on campus doing things like human computer interaction or machine learning. Most of them work for 5-6 hours a week during the school year, usually coding models or writing scripts for a large team. If the group obtains a result, they often get to be 5th or 6th author on a publication or listed in the acknowledgements section.   

Which route would you recommend for an optimal grad school application? 

Thanks!

[Ryuk]

I got into a few top CS PhD programs as a more theoretical applicant, but I am FAR from an expert.

I had a similar research "experience" early on in my undergrad. I was given some data and a few papers, then told to find something interesting. It went nowhere. Obviously, it would be extremely impressive if you were able to produce concrete results as an undergrad with so little guidance. This would almost certainly be better than any lab work. However, it might not be realistic for you; it certainly wasn't for me.

Are you applying this fall? If not, REU programs are pretty much made for your situation. The professors in an REU will ideally teach you to swim instead of throwing you in the pool like your 2 current profs. Researching is a skill that takes many years to master and is pretty much the whole point of a PhD. These summer programs are free to apply and actually pay you to attend, so there's virtually no excuse.