Good careers in Computational Physics?

[chi-young]

I'm trying to go into big data and I like physics so I'm going for a degree in computational physics. I'm thinking of learning Hadoop and Storm, because all of the corporate careers seem to require them. Am I barking up the right tree? What kind of career path will a Bachelor's in Computational Physics prepare me for?

 

[TomServo]

I think big data would prefer CS majors who study things like searching, sorting, graph theory, algorithm analysis, as opposed to the numerical methods kind of programming physicists do.

 

[gbeagle]

I used to work in 'big data', and I agree with what TomServo said. It is very much a different sort of problem than the type you work on in computational physics. The work in 'big data' more closely resembles experimental physics: Collecting data, analyzing it, and trying to draw insights/conclusions from it.

 

[StatGuy2000]

In addition to the CS majors pointed out by TomServo, companies and organizations working in "big data" will tend also to prefer CS majors who have studied machine learning methods (e.g. neural networks, K-means clustering, decision trees, etc.).

Since "big data" involves inference on data by definition, statistics majors (who have earned either a MS or a PhD) are also in demand.

 

[kinkmode]

I just interviewed for a big data company. I think I did reasonably well even though I didn't get the job. While many people working there were from CS backgrounds, there was a small group in the Informatics department called Analytics. Not one of the four people in it came from a CS background. They didn't seem overly concerned about the lack of experience with the specific technologies (Hadoop, MapReduce, etc.), though I'm sure that would have helped. I was able to puzzle through their 'technical' questions reasonably quickly since they didn't rely too much on specific knowledge of a field.

Note that I am/was an experimental physicist whose done my fair share of modeling along the way, but mostly a lot of data collection and analysis. I also have a Ph.D. The head of the group also had a Ph.D., but in Applied Math.

I would also like to note this obvious point about technology. It's a quickly moving target. Hadoop hadn't even been invented and 'big data' wasn't a thing when I was in college. Heck, Hadoop came around past the half way point of my graduate experience. I think Hadoop and the like are good skills to have for the market NOW, but who knows what it will be like in 5 or 10 years.

This is just my take based on my brief experience. I'll also just throw out there that most of my friends and associates who specialized in computation physics, albeit at the Ph.D. level, are all gainfully employed and seem to have less trouble then the experimentalists finding work and transitioning to other fields.

 

[TomServo]

I'll also just throw out there that most of my friends and associates who specialized in computation physics, albeit at the Ph.D. level, are all gainfully employed and seem to have less trouble then the experimentalists finding work and transitioning to other fields.

Good to know! What sort of jobs do they do? Computer graphics? Finance? Energy? Sushi chef?

Also, what counts as "specializing in computational physics"? Doing your thesis in group that is explicitly about computational physics, or does this include, say, theorists who do loads of programming?

 

[kinkmode]

Most of them do physics research in nuclear fusion or weapons related work. One does more astrophysics-related simulations, though dabbles in the fusion/plasma physics world still, and one does simulations related to medical physics/nuclear medicine. I do have some friends who do high frequency trading finance, but I think they were experimental physicists in grad school.

None of the people I'm thinking about have really left research except for the finance guys, so no one develops iPhone apps or anything.

'Specializing in computational physics' => Most people in my field whether they are experimental or theory have to do some computational work along the way. I didn't mean that. I meant people who essentially only do numerical/computation work and not that much analytical theory or experiment. That's a pretty big class of researchers in fusion. That's not to say that if you were a theorist who was very proficient at computation you couldn't also make the transition. It's just that if you really like analytical work, most of these people I'm talking about don't do much of that.