I have good experience in the field of Deep Learning. I have been using Tensorflow for more than 8 months now. Few of the projects that I have done in the field of Deep learning are:
1. Classification of the non-MNIST dataset. I used various approaches such as Logistic Regression, Convolution neural networks etc. I achieved an accuracy of around 96%(highest is around 97%).
2. Prediction of missing word in the sentence. For this, I implemented 2 approaches, word2vec + CBOW and bidirectional LSTM.
3. Currently, I am working on image captioning system.
I am quite sure we can use word2vec and may be LSTM for your project. Please tell me more about your project like what is dataset etc.