MQL 5 AI trading bot

I am looking to develop an AI trading bot in MQL5 with tensorflow for scalping

Every tick :

- the information must be sent to the strategy pool :

* in this pool we have to develop different indicators to start with we can start with the 10 main ones and then we will add more to improve the system

* at the output of each indicator we must have an output if we must buy, sell or do nothing: 

At the output of each strategy, the information must feed a neuron at the input (each neuron can be processed by openCL), the neural algorithm is LSTM with Adam (Adaptive Moment Estimation Method)

with the learning on the history (for each strategy and the output) we should be able to develop an AI ==> in output we will obtain the signal to do: buy, sell, do nothing

and to know the positioning, another module for money management, SL, TP