Water is one of the most precious resources for the human population and could easily become scarce with the emergence of climate change phenomena which is greatly affecting the world. Preventing and reducing water losses and general preservation of water should be a priority in order to ensure that it doesn't become a scarce resource as it plays a vital role in both the human life and within a country‘s economic development. Water losses in urban water distribution networks are a common occurrence which are usually caused by worn out pipes, corrosion, high pressure, water hammers etc. It is extremely important to detect and localize a leakage location in a water distribution network in order to stop it and remediate the damage. Within this project, we propose a data-driven methodology for leak detection and localization which is based on artificial intelligence, more specifically, a hybrid of machine learning and swarm intelligence optimization algorithms. Data-driven methodology such as machine learning has been proven to work well on water distribution network related problems such as leak localization and contamination source detection, however, the approach still suffers from a lack of robustness. To improve on this, it is essential to create a hybrid approach which uses machine learning methods to detect anomalies such as leaks and optimization algorithms in order to localize the leaks in a water distribution network. This hybrid methodology will be implemented on the BURA supercomputer of the University of Rijeka in order to maximize the computational efficiency and increase the robustness. The leak scenario data will be generated using the previously validated water distribution network hydraulic simulator. Finally, the proposed methodology will be investigated on real data obtained from the Water supply and sanitation department of the city of Rijeka.