Foreword: This is not an idea, and what I'm going to talk about is mostly my life, I wouldn't give two craps should I be you reading ne right now :D
Dear fellows traders,I'm in a situation I'm not really sure how to handle.Because it's about trading, cryptos, and money, I figured posting it here would not be completely absurd.
First of all, I'm not trying to sell you anything (admittedly, that may be an option at some point. But first, I'm a coder, I'm good with code, I'm really bad at selling stuff to other people. And second, I wouldn't trust a total stranger with my claims. From your point of view I'm a random dude on the internet. "Random person on the internet" should be given about the same level of trust as "Convicted pedocriminal applying for a child custody position")
I'm a machine learning researcher. For the past 15 months, I've been working full time on a trading bot.I won't tell publicly what my key findings are, only: I'm using state of the art techniques of several fields. My current backtests, (with huge fees to take slippage into account) average around 5% a day, with little to low variance.
Some of the principles my bot is based on:
* Turning order books movements into a sound resembling signal, and using state of the art models for sound processing.
* Crafting indicators with genetic programming
* Figuring trendlines, patterns, key levels (which is hugely more tricky that I could have thought before starting that project)
* Elliotish waves (I'm not actually using elliottism, rather, I turn market signal into waves, turn waves into matrixes, and use state of the art sequence prediction models)
* Training models without gradient descent, allowing Reinforcement learning but thousand fold faster converging
Problem is: It's not currently ready to trade. Because putting together complexes interacting software bricks is no piece of cake. I need to write tests, and make a lot of adjustment to take into account differences between backtests.
I'm currently looking for a coder job. Because I don't have time to trade, and those 15 months have been financed by my savings, which is running dry.
I need between 2 and 5 months to turn my project as it is into what I believe will be a money printer.
You don't have to believe me (again, random dude on the internet) nor trust I'm not delusional. I just want opinions on what to do next. Or ideas.
Here options I've thought about:
* Finding a job, and keep the bot as a side project
* Make a "fine" version quickly and allow a limited amount of people to use it, with subscription and a neat interface
* Send job application to existing platforms (as Mudrex), propose them to share my code and be paid for that.
Not sure what I'm expecting from this post (trolls, I do expect trolls :D ). But plenty of smart folks are there, it may worth a try...
Dear fellows traders,I'm in a situation I'm not really sure how to handle.Because it's about trading, cryptos, and money, I figured posting it here would not be completely absurd.
First of all, I'm not trying to sell you anything (admittedly, that may be an option at some point. But first, I'm a coder, I'm good with code, I'm really bad at selling stuff to other people. And second, I wouldn't trust a total stranger with my claims. From your point of view I'm a random dude on the internet. "Random person on the internet" should be given about the same level of trust as "Convicted pedocriminal applying for a child custody position")
I'm a machine learning researcher. For the past 15 months, I've been working full time on a trading bot.I won't tell publicly what my key findings are, only: I'm using state of the art techniques of several fields. My current backtests, (with huge fees to take slippage into account) average around 5% a day, with little to low variance.
Some of the principles my bot is based on:
* Turning order books movements into a sound resembling signal, and using state of the art models for sound processing.
* Crafting indicators with genetic programming
* Figuring trendlines, patterns, key levels (which is hugely more tricky that I could have thought before starting that project)
* Elliotish waves (I'm not actually using elliottism, rather, I turn market signal into waves, turn waves into matrixes, and use state of the art sequence prediction models)
* Training models without gradient descent, allowing Reinforcement learning but thousand fold faster converging
Problem is: It's not currently ready to trade. Because putting together complexes interacting software bricks is no piece of cake. I need to write tests, and make a lot of adjustment to take into account differences between backtests.
I'm currently looking for a coder job. Because I don't have time to trade, and those 15 months have been financed by my savings, which is running dry.
I need between 2 and 5 months to turn my project as it is into what I believe will be a money printer.
You don't have to believe me (again, random dude on the internet) nor trust I'm not delusional. I just want opinions on what to do next. Or ideas.
Here options I've thought about:
* Finding a job, and keep the bot as a side project
* Make a "fine" version quickly and allow a limited amount of people to use it, with subscription and a neat interface
* Send job application to existing platforms (as Mudrex), propose them to share my code and be paid for that.
Not sure what I'm expecting from this post (trolls, I do expect trolls :D ). But plenty of smart folks are there, it may worth a try...
Comment:
PS: The chart you see is made of trendlines that came from my bot :)