DeepSpeech is an end-to-end architecture where deep learning replaces traditional hand engineered speech to text algorithms. The model performs well, independent of speaker adaptation as it directly learns from data. We will take a brief look at the model architecture of DeepSpeech.