is it harder to get a trading developer job in ny than chicago

Discussion in 'Automated Trading' started by kci, Jul 22, 2009.

  1. kci

    kci

    For junior developers?

    It seems like in chicago a junior developer will almost never get a chance at a trading job with the requirements asked for...

    Is it like that in NY?
     
  2. I am doing well with job searches in NY and Chicago, and I am somewhat junior in the sense I did research in academia and not real work.
     
  3. rosy2

    rosy2

    what can you do? I assume "junior" just means you cannot do anything
     
  4. kci

    kci


    By junior I'm saying 1.5 to 2 years of experience in c++
     
  5. rosy2

    rosy2

    again. just because you have 1.5 to 2yrs experience doesnt mean anything. i know people with 10+ years and they can't do anything. Others are straight from school and are incredible.

    what can you do?
     
  6. kci

    kci

    I feel the same way as you that some people come straight out of school and are very good and yet some people that have experience are not...

    I am very good in C++, Java, C#, SQL.... However my observation is it seems like in the chicago job market for trading developers if you don't have 5 years or so experience, you can't even get an interview and probably won't even meet the criteria to apply for a position. Thats why I ask if its the same way in NY...
     
  7. rosy2

    rosy2

    very simple.

    post a link to some code you wrote and why you wrote it like you did
     
  8. kci

    kci

    A better way to test my claim is to show me a chunk of code and ask me questions about it or different ways to implement the same thing etc

    Asking me to post you a link means I can just show you someone else's work. I'm game for any code challenge in c++, java or c# only.
     
  9. Great... we have another "straight out of college idiot" who doesn't know who he's talking to.

    So here's a question:

    The attached code is the user-side implementation setup for a broker I used for Forex using QuickFIX that I had 3 years ago. I was able to reduce the code to about 1/5. Tell me how I can tweak the OpenSource QuickFIX code to reduce this to 1/2....

    (At this point... I haven't touched the QuickFIX source... but I tweaked the source so that I can drastically reduce the code without influencing other setup codes for other brokers...)

    Please post the QuickFIX code (there's a few) and the result C# code corresponding to the tweak.
     
  10. nitro

    nitro

    Gann,

    What you are asking is unfair. You have had years to think about the problem.

    Also, I looked at your code. Some of your QF code looked hacked, not engineered. One suggestion is to surround all your comments with #region OldImplememtion #endregion. That would make your code easier to read. Even that is probably messy, since if you are using source control, and good comments in the checkin, you can go back and look at why you did things the way you did, and changes you made, there. You can always tell programmers that either work on one man teams, or are egocentric and don't care about the next person tha has to maintain their code. Code should be a pleasure to read. Like a novel.

    Finally, optimizations to QuickFIX in my experience (which is limited since I don't use QuickFIX) is not to optimize the client code, but to optimize the communications module.

    Things like this are horribe:

    Code:
    [b]public void sendNOSMarket(string symbol, string side, int size, long OrdID)[/b]
            {
                string[] currency;
                [b]currency = symbol.Split('/');[/b]
    
    or
    
                    if (mdupdateaction.getValue() == [b]'2'[/b])
                        [b]Console.WriteLine("Enter");[/b]
    
    
    Why would you hardwire code like that in an FIX client? What does 2 mean? Why are you passing four parameters instead of an object of type OrderParameters (which would then allow you to inherit for each case where the symbol representation changes)? Why use Console.WriteLine that go away when code is compiled release? Use Tracer and listeners and then you can log errors no matter what.

     
    #10     Jul 23, 2009