Sentiment analysis in deep neural network

Dear Bidders, I am looking for a resource who is comfortable with deep learning, especially Autoencoders and BiLSTM on Text. It is a sentimental analysis project. I already have dataset and code ready BiLSTM. I also have code related to Autoencoders. I shall provide you with that. However, the requirement of the project is to extract features using Autoencoders.. and to use those features as dataset to be learned upon using BiLSTM. For BiLSTM, you can use the same code. For Autoencoders, perhaps, a few steps could be taken from the jupyter notebook code file. I have also all the pre-processing steps included in the code. The only requirement is to extract features from Autoencoders and apply it to existing BiLSTM code in such a way that accuracy is good as possible. Please put some concentration on increasing the accuracy :)