Advice for an applied math PhD Student

[greatdane]

Hi guys,



 



I have been a long time reader of these forums and I have benefited a lot.  After searching through old posts I think I have a question which hasn't been directly answered, partly because it is very specific.  I am beginning a PhD in applied math, and there are two research areas I am choosing between: pattern theory/computer vision and theoretical, proof based work with PDEs (mostly analysis).  Although they are rather different, I am very interested in both and it has been a tough choice.   I really enjoy analysis but many of the applications of pattern theory/machine vision interest and excite me.  I think I might like to go into finance after completing the PhD, so I am trying to factor into the decision the effect of this choice on a career in finance.



 



Pattern theory is much more applied than the analysis research, for instance, it would involve some programing (mostly MATLAB) and working with real data.  In addition, pattern theory involves significantly more statistics, although it still it involves some PDEs and differential geometry (but I am sure to a much lesser extent).  I am trying to determine if, in general, employers would prefer someone who has a very strong analysis background but very little experience with real data and computation, or if they would prefer someone with a decent PDE/analysis background and more applied experience, as well as statistics knowledge.   For whatever it is worth, my ideal quant job would be a stat arb type role, coming up with trading strategies and algorithms.  This seems to lend itself well to pattern theory, but I have heard that these jobs are very hard to come by, so perhaps I should be more concerned with the skillset to get a more typical quant role. 



If it doesn't really make a difference, or it is impossible to tell, I'm sorry for wasting your time.  I just thought I should check with a knowledgeable group of people to see if this choice would greatly impact my finance job prospects.  Thanks in advance for any advice you are willing to give, and for providing a lot of information to me through these posts over the years.

[bloodninja]

Do both. Probably sounds crazy. But, I can think of a couple  reasons for contemplating the attempt of both.



1) It's a hedge if you do both. For example, the old saw about vocation and avocation. Yes, it is realistic and sensible to take a pragmatic view and pursue a subject because it has a marketable future. However, doing so, as sensible as it is, can also be drole. Jack the dull boy etc. So, hedge the burnout of sensibility with something that is more interesting and done for the sake of doing it.



2) By doing both, there is a chance that one subject can either inspire the other or directly connect to it. True, the chance of explicit connection is slim, but if you can connect the two...well then that might be very interesting. More probably, by doing both when you're stuck with one then you can always go back to the other until 'inspiration' occurs. Whichever finishes first, finishes first. Along these lines you could consider either serial, sequential / piece wise,  or parallel approaches to the two projects. Alternately, one project could be the day job and the other could be the 'hobby' / avocation



Having worked night and day jobs in disparate fields;  and pursued vocations and avocations and seen them flip back and forth and contribute to one another I am inclined toward 2....along the lines of mathematics being the art of connecting seemingly unconnected domains.



best of luck



hih

[mj]

i did my pure maths work in theoretical PDE and whilst it's rarely been directly useful, the intellectual training certainly has been