I am a Postdoctoral Research Scholar in the Department of Electrical Engineering at Stanford University where I work with Andrea Goldsmith. My research focuses on developing novel data-driven techniques based on machine learning and statistics for design and analysis of communication systems in settings when accurate and simple models are not available. In my work I have developed new tools and techniques for multiple aspects of communication system design using machine learning, including: novel detection algorithms that dynamically adapt to changing channel conditions and neural-network based joint source-channel encoders and decoders that can be applied to text, images and video. I have also applied this approach in the emerging paradigm of molecular communication, where chemical signals are used for information transfer. Some of the techniques I develop can be adopted to problems in other areas such as speech processing and natural language processing.