Data and extracted knowledge have become a crucial part of the digital transformation of many sectors. The process of extracting knowledge from data relies on data mining, which requires optimisation for efficient and specific data collection. However, data mining cannot cope with situations where data characteristics are extreme. This creates a developmental bottleneck, and solutions are needed to allow for efficient data mining across computing continuums that can handle extreme data characteristics. The EU-funded EXTRACT project offers a solution to this challenge. It will develop a data-driven open-source software platform that will utilise a vast array of computing technologies to guarantee safety, improved performance and energy efficiency while allowing for extreme data mining.