Custom machine learning research platform

Discussion in 'App Development' started by fan27, Sep 4, 2017.

  1. wintergasp

    wintergasp

    Where are you going to find the first 50m$ for your fund ? If your answer is anything but "I have a seed partner that can write a check for that amount" then forget about it. Running a fund has extremely high fixed costs and this is probably the worst time out of the past 25 years to launch a hedge fund (in terms of investors interest, regulation etc ...)
     
    #11     Sep 5, 2017
  2. Simples

    Simples

    Congrats on reaching a milestone. Will you open source it?
     
    #12     Sep 5, 2017
  3. fan27

    fan27

    I might open source portions of it. For example, my package for loading market data is currently available:
    https://github.com/fasterbull/marketdata

    I might also make available the backtest package I am currently working on. Will keep you posted.
     
    #13     Sep 5, 2017
  4. fan27

    fan27

    I don't disagree. Seems like the only realistic thing to do is to prove out my technology via actual trading results and then see what avenues of opportunity open up.
     
    #14     Sep 5, 2017
  5. wintergasp

    wintergasp

    The reality is that the hedge fund game is not about performance but about marketing. Even if you show great numbers, you still won't attract (enough) capital.
     
    #15     Sep 5, 2017
  6. fan27

    fan27

    I have given this topic some more thought. Seems like I would have a better chance at developing my platform to be used in some sort of SAAS product. I think other traders would find it useful. Rather than a monthly fee I could charge per backtest where a backtest would consist of testing out thousands of permutations of features. I am not aware of any platforms out there currently that can easily do that without being terribly slow. Stay tuned!!

    fan27
     
    #16     Sep 21, 2017