We consider the problem of joint source and channel coding of structured data such as natural language over a noisy channel. The typical approach inspired by information theory to this problem involves performing source coding to first compress the text and then channel coding to add robustness while transmitting across the channel; this approach is optimal with arbitrarily large block lengths for discrete memoryless channels. Given documents of finite length and limitations on the length of the encoding, we achieve lower word error rates by developing a deep learning based encoder and decoder. While the information theoretic approach would minimize bit error rates, our approach preserves semantic information of sentences by first embedding sentences in a semantic space where sentences closer in meaning are located closer together, and then performing joint source and channel coding on these embeddings.