Trying to decide if it makes sense to get a Masters in something CS related-- if so, trying to decide between Stats & CS.

[jwdink]

Hi Folks,

 

I'm currently in a PhD program for cognitive psychology. Academia's tough, and it would be good to have some marketable non-academia skills, as a backup. Luckily, I'm pretty good at programming, get a good amount of practice with it, and really enjoy it. I also find Machine Learning, algorithms, and AI really interesting (which makes sense, given my interest in cognitive psychology).

 

So it'd be nice to have the option (if needed) of moving out of academia after I graduate, and into a career in industry, focusing on ML/AI. How can I best set myself up for this?

 

Here seem to be my options:

• Focus on the PhD, try and make research more computational, try and work on CS-y personal projects and build up a portfolio of these sorts of things.
• Continue with the PhD, get a masters in stats meanwhile. This seemed like a good option at first. The stats department at my university offers a program for people who are getting a PhD in another department-- for me, just take 8 courses in their department, no thesis even. I have 8 quarters left, lines up nicely. I know that machine learning and stats are heavily related (though it's unclear to me whether AI and stats have much relation). HOWEVER, I missed out on the intro sequence for this year, and will have to wait until next year to take that. Courses that don't depend on this intro sequence are either stuff I'm super familiar with (linear regression) and/or stuff that has nothing to do with what I'm interested in (e.g., meta-analysis). In addition, the director for this program has been very difficult to pin down to talk to, making it difficult to discuss alternative options with him.
• Continue with the PhD, get a masters in CS meanwhile. I would love this-- the courses are explicitly on the topic of my interest (machine learning, AI), and I enjoy programming more than I enjoy math. However, this department doesn't explicitly have a program in place for people in other PhD programs (like the stats department does). Instead, I'd be taking the usual MS requirements, which are a bit heftier: 12 courses OR 9 courses + a thesis. This would be a lot to juggle with my PhD, might be difficult to get my advisor to approve of, etc.

So my questions are:

• Would the stats degree really help get me into the areas I'm interested in? I've heard it's relevant for ML, but for some reason I'm second-guessing that.
• Will potential employers want to know that I've completed a thesis/ master's project anyways? One of the downsides of the CS is that it's quite a bit more work-- but if this is the sort of thing that employers are going to be looking for anyways (as opposed to "oh yeah I took 8 courses"), it may be worth it. Or can I be pretty vague with employers on how I got the MS? This is related to the concern that I'll be coming out of this program without any CS internships or the like-- just book learnin'.
• Are either of these Masters actually helpful? Lately I've been hearing that most ML jobs require a PhD anyways. Now, technically I will have a PhD-- it'll just be in cognitive psych. Does this help?

(Bonus Question: Can someone describe what "Statistical Computing" is-- or rather, what a typical class on "Statistical Computing" is like? This is one of the classes I could potentially take this year. It seems like its more appropriate for my interests than e.g., "meta-analysis," but I don't know if it'll assume lots of background/foundational knowledge that I'm missing out on by not taking the introductory series this year.)

 

Sorry for the the length of this post. Any help would really be appreciated, I'm feeling pretty paralyzed. I think my heart says CS, but my head says stats.