Scheduling optimisation through machine learning methods The EU Amazon Transportation Services team is building an application that will help the operational division in making better scheduling decisions. There is an existing application, called CODA, that is able to perform different tasks, such as predicting the demand and truck fill-rates or estimate dynamic transit times for the fortnight to come. The argument of this thesis will be to improve or even replace the existing jobs' logic while adapting them to new zones with different data and context, such as India and NA. The techniques used to do so will range from implementing new ML models, use statistical analysis and data science methods to improve the application while also working on the application logic itself (mainly backend and data warehouse design).