Coffee is one of the most commonly consumed beverages in the world. Coffee extraction is an established art and business, practiced all over the world. In addition, the coffee supply chain unfolds on a global scale and the coffee market is inherently international in character, with an estimated average annual growth rate of 1\%-2\% for the next 10 years. Players in this sector are therefore constantly looking for product and service innovations to increase their competitiveness. Such coffee fame and the search for innovation have stimulated extensive coffee-oriented scientific research, an important component of which is in-depth knowledge of the physico-chemical processes occurring in the coffee preparation. Another goal of the coffee industry is to customise the coffee beverage in terms of both taste and health. The latter problem has attracted the interest of the scientific community, as it is an intriguing inverse problem that can be tackled with a hybrid approach leveraging tools of Computational Fluid Dynamics and Machine Learning. In this work, we present a solution of the taste customisation inverse problem of espresso coffee through simple Machine Learning techniques. The procedure was tested with an extraction campaign and a panel test with experienced tasters. The results obtained are promising, so this tool, refined with further validation, can be valuable to the coffee industry.