Careers for someone who mostly knows about scientific computing?

[davidbenari]

I study Physics and my major requires me to take 6 optional courses in any area of my interest.

I am interested in pursuing a career as a programmer, but my university doesn't allow me to choose Computer Science courses due to their high demand. Perhaps I can take one or two CS courses but that's about it.

What I can take are courses in the Computational Science area. These courses are mostly in what you would consider Scientific Computing. For instance, there are courses in Numerical Methods for PDEs, Numerical Fluid Dynamics, Finite Element Methods, Statistical Computing and Learning, Advanced Data Analysis, Algorithms, Parallel Algorithms etc (you probably get the idea).

These courses use C++, Python, Fortran, R, and some of them, like Numerical Fluid Dynamics allow me to program in a supercomputer and work with Linux (I don't know why the course requires Linux knowledge).

Besides the fact that these courses are interesting in and of themselves, I was thinking they could provide additional knowledge in programming and help me become a programmer.

There isn't much of a market for numerical methods (according to my google searches). So my question is: will the things learned in these courses be transferrable to some programming job, perhaps as a Quant, to say an example?

What sort of jobs would fit this type of knowledge? I'm not asking you to mention a job that, e.g., is specifically about solving PDEs since that would be quite limited. But maybe you could tell me jobs that have the same intellectual atmosphere. Preferably, do not mention jobs that require Masters or PhD, which is what most Quant jobs ask for AFAIK.

Any thoughts? Am I wasting my time in these courses if I want to be a programmer? I know I should've studied CS if I wanted to be a programmer, but the fact is I did not and chose physics instead.

By the way, I already know programming and have taken courses in computational physics and numerical methods.

Thanks.

 

[Dr. Courtney]

In real life, you got to go one semester at a time with a long term guide as a plan, but the courses you want long term may not always be offered.

It is usually easier to advice the next semester than the whole degree.

 

[davidbenari]

I'm about to finish my degree by the way. These 6 courses come after I've seen all physics and general education courses. I've already added 2 of those courses for next semester, although I can change them. I added Numerical PDEs and Numerical Fluid Dynamics. I was still worried about the mentality with which I am adding them (wanting to gain programming skills), instead of having an actual passion for those two subjects.

I don't think its unrealistic to plan a career as a programmer right now. I hear of many physicists becoming programmers and my guess is that all they know is scientific computing and not the standard core CS courses (like data structures, algorithms, databases, etc).

 

[rootone]

It could be a good idea to register with a number of IT temp staff recruitment agencies.
It worked for me, although some of the jobs can be dull at first, (stock market algorithms and bleh), it is a career path

 

[davidbenari]

It could be a good idea to register with a number of IT temp staff recruitment agencies.
It worked for me, although some of the jobs can be dull at first, (stock market algorithms and bleh), it is a career path

I've heard of programming "boot camps", you mean those?

I actually find financial programming interesting. How did you contact those guys? Did you require graduate studies?

 

[mpresic]

I have not been in a job search for 2-3 years, however you stated many observations which are counter to my observations.
First, the job market for numerical methods is not in decline. Although there are sometimes numerical libraries already developed, some organizations do not want to pay the rates for software licenses and are looking for people who can develop their own algorithms to replace those algorithms to which they currently need licenses. Even in the case where numerical algorithms have been developed, physicists can sometimes develop check cases to verify the algorithms are being used correctly, numerical precision is not being lost, and the results make sense.

Second, many of my colleagues who I work along side with do know data structures, algorithms, and databases, although I do not think they learned it in school, but learned it on the job. It is not true that "all they know" is scientific computing. Indeed many of my colleagues involved in data structures and databases are not as knowledgeable about scientific computing.

The UNIX operating system is widely used, and than is why Linux experience is so valuable for C++, fortran, python, R and other programing languages.

It is good that you have courses in numerical methods, and computational physics. This may be your best feature. If your google search suggested you cannot get a job in numerical methods, I wouldn't trust it. By the way after I got hired, I found out that all the advice I got regarding what computer languages were out-of-date, and which were hot was misguided. Sometimes experts in old languages like COBOL or FORTRAN are needed because a lot of old code needs to be transitioned or understood, and the developers/originators have long since retired or otherwise unavailable.
Even at job fairs and career counseling centers, you can get a lot of conflicting information as to what skills are needed and who gets hired. Job seeking takes a lot of effort. (I found it harder and more time intensive than getting into grad school). There may even be substantial luck involved.