Serving the Quantitative Finance Community

 
User avatar
street
Topic Author
Posts: 0
Joined: March 27th, 2008, 1:51 am

Tech Route.

September 15th, 2009, 3:27 am

Anybody here more involved with Tech Side of the market.By tech side I mean systems development, message passing middleware, data feeds etc etc.It seems that this the direction my Academics are leading me. If you had to pin down my field its 'distributed event based systems'. Lots of queuing stuff, distributed query planning shit like that. I have a decent enough finance background I've read the right books, I have also taken most of my universities courses related to QF. Plus, all the regular eng classes that might be useful stochastics, optimization, numerical programming etc.What organizations should I be thinking about? My C++ skills suck. As I spend most of my time in MS-land. To be honest if I was going to learn something completely new, I would look at programming gpu's or fpga's.Recommendations?
 
User avatar
DominicConnor
Posts: 41
Joined: July 14th, 2002, 3:00 am

Tech Route.

September 15th, 2009, 1:14 pm

GPUs are certainly warming up, and if you are not risk averse I see them as a good bet for someone with your set of skills, certainly better than FPGAs which I see as not growing much.There is a mechanism in the job market that I've not really explained before, which requires that a skill be both rare but wanted quite strongly.By "strongly" I don't mean aggregate demand like for Java programmers, but specific employers who want it, and will not accept substitutes that are "near".Of course a skill can be in both sets, but GPUs are never going to be a mass market programming interface.In this context employers will focus on the critical skill and be more forgiving of you not fitting their other requirements (usually).Of course this also pushes up money, but with greater variance and lower average than you might think. Employers do not know how to price your skills and frequently think they may be being ripped off, so handling this is a delicate operation.Critical transition skills always exist, the trick is to spot them before they become common or obsolete (or both).Thus if street got his GPU skills to a good level before the masses then he has a good chance of nearly getting jobs nearer to what he wants....but he won't.He doesn't know C++, and for some reason seems to believe that has something to with being a victim of Windows.I can't imagine even trying to place a GPU quant developer who can't do C++, or as a theoretical minimum, very good C99Street has no excuses for this, modern Visual C++ is a highly standard version and is downloadable for free together with the best environment for learning C++ known to man, Visual Studio.Most Windows programming interfaces are set up for C++, as is the case for the Unixes.The current leader in the GPU space is CUDA which is a variant of C99, and although one can use Python or even Matlab, the grownup way forward is in the C++ family.Java is a non starter and although I've read of work for C#, frankly it looks as promising as VBA.If Street takes on this appallingly tough task, and does well, then there are several types of firm out there who might employ him.Algo trading seems to beckon, GPUs are nascent in this space and someone who can slice down to the market feeds should be employable.Some of the large banks (no I'm not saying who), are re-coding parts of their analytics libraries for GPU
 
User avatar
street
Topic Author
Posts: 0
Joined: March 27th, 2008, 1:51 am

Tech Route.

September 15th, 2009, 5:21 pm

Thanks DCFC. I was hoping you would chime in. You have lit a fire under my ass.always more to learn. Anybody else?