Transportation fuels like gasoline and diesel are a cocktail of several thousand individual molecules which makes understanding their chemical composition even more challenging. The present talk proposes and tests the following hypothesis, “Can we use a fuels immense molecular data to predict its properties using AI techniques?” The talk will take a deep dive into various fuel properties like octane number, cetane number, sooting propensities, flash point etc. and the development of their prediction models using machine learning techniques like neural networks. The application a novel functional group approach for designing future fuels using ML/genetic algorithms will also be discussed.