Machine learning is a key component underpinning the recent successes of intelligent systems and data analytics engines. Machine learning algorithms trained on data can perform impressive tasks, but often at the expense of massive computational resources. Reducing these requirements is the aim of the EU-funded SLING project. SLING is developing a new generation of resource-efficient algorithms for large-scale machine learning solutions that can be easily applied to real-world scenarios. The solutions developed within the project make machine learning more accessible and sustainable, significantly enhancing the prospects for developing truly scalable intelligent systems.