switching from Science to CF

Joined
6/25/12
Messages
3
Points
11
Hi all, I am new here, looking for some advice.

I have a BS in Physics/Math and a PhD in computational Chemistry, I am doing a postdoc right now, also in a computational field. I am looking into switching to finance. All research projects that I ever worked on were related to modeling, statistics, dealing with huge quantities of data, etc. I am not a very good programmer (naturally, I concentrated more on science rather than on programming), but I have some C++ and Perl skills, that I believe I can improve with some effort.

So, here are my questions. Do I need a degree in computational finance? I.e. how important is that to have a formal diploma in order to find a job for someone in my situation?

Also, do you think I can take a few necessary courses or is it better to go through a full program? I currently live in Pittsburgh and now I am looking into MSCF program at CMU. I am just trying to explore my options at the moment and figure out whether I need to go to school again.

Thoughts? Comments? Thanks in advance!
 
You would have a hard time finding jobs if you don't know which ones to apply to and how to prepare for them. You will be asked on brain teasers, programming questions so they need to be addressed.
Go through the list of books/guides on this list
If you can find some good recruiters who willing to help you, then try to get some leads and interviews from them.
 
Thanks a lot for reply and for the link! I will certainly study this.

Yes, I realize that I have to prepare, I am trying to figure out how far do I need to go with this.
 
Ideally, you go through the list and pickup the knowledge and skills necessary to interview successfully. Then you will avoid spending the next few years to prepare for MFE admission, studying, etc and 100K cost.
You don't know if you are ready until you get out there. The market for entry-level market is really soft right now. Risk/regulatory will be a major hiring sector.

Or if you want to get a job the fastest way, do as advised by a hiring manager here
https://www.quantnet.com/threads/learn-sql.7554/
 
One suggestion would be to take a very involved project from one of the books or of your own idea and code it out yourself in a base language without calling any libraries. If you are motivated to finish the project, this is a good way to gain proficiency in coding (and make something cool for yourself or to show off to whomever).

Doing shorter, discrete exercises in programming has less value added in my experience because you tend not to integrate the knowledge as you would when you have to continually revisit / debug all the moving parts of a more involved project.
 
Jose,

This is an excellent suggestion and something I have been hunting for some time. Would you be willing to point me to a source (book/website) that has such a serious project? Thanks.
 
Here's one thought: you can pick an R package or packages that you find interesting and useful, familiarize yourself thoroughly with their functionality and try to recreate everything that package or packages do in a C or C++ written application with a GUI. Again without using libraries in the sense that you don't outsource, for example, matrix multiplication or finding eigenvalues; however you will probably need a basic math library for things like logarithms, powers, square roots, etc.

I'm still in school myself and don't necessarily know what would be most relevant you. One thing I have an academic interest in are state space models. I guess if it were me, I would try to deconstruct something like these packages:

http://cran.r-project.org/web/packages/dlm/vignettes/dlm.pdf
http://cran.r-project.org/web/packages/dlmodeler/dlmodeler.pdf
http://cran.r-project.org/web/packages/sspir/sspir.pdf

All of these packages include other packages, so you would have to go kindof down the rabbit hole to figure out what was going on at a low level. I haven't tried it myself, but I'm guessing that recreating these in a base language would be rather involved.

I'm a little more bent to statistics than probably others. Maybe there is some other project directly involving something like derivatives pricing that someone could recommend.
 
Back
Top Bottom