I have one strategy that is hard to implement, at least for me because I'm not very familiar with computer science. It requires strong math knowledge.
I'm looking for someone who is familiar with crypto trading, python and data science.
Could you implement AI based strategy when input for neural network is tens of indicators calculated for each candles updates (5m) and network can auto-learn from past candles and correct itself?
And all this should be using SVM algorithm like described in following documents.
You can write your custom code or even better - use freqtrade or any other good framework to save your time and concentrate on algorithm itself.
1) [login to view URL]
2) [login to view URL]
3) [login to view URL]
If you read this carefully and think you can afford it - write "I'm in" in the beginning of your application.
Otherwise I'll just filter your application out.
It should be final working solution using binance api and able to work in a demo mode (with virtual orders) which is supported by default when you don't use binance credentials.
If it's possible, strategy should use freqtrade or at least any other available trading framework.
But if framework doesn't support structure of AI algorithm let the algorithm be based on minimal code structure which works with binance api and proves efficiency of algorithm on a demo data.
Good plus is ability to test it on historical data
I can help freelancer with building indicators data based on candles and np/pandas python packages.
Default freqtrade strategy coming with installation already contains all examples.
So indicators are not an issue.
The most complex part of the project is building AI structure and using those indicators properly with good learning weights