Climate change is one of the most urgent problems facing mankind. Implementation of the Paris climate agreement relies on robust scientific evidence. Yet, the uncertainty of non-greenhouse gas forcing associated with aerosol-cloud interactions limits our constraints on climate sensitivity. Radically new ideas are required. While the majority of forcing estimates are model based, model uncertainties remain too large to achieve the required uncertainty reductions. The quantification of aerosol cloud climate interactions in Earth Observations is thus one of the major challenges of climate science. Progress has been hampered by the difficulty to disentangle aerosol effects on clouds and climate from their covariability with confounding factors, limitations in remote sensing, very low signal-to-noise ratios as well as computationally, due to the scale of the big (>100Tb) datasets and their heterogeneity. Such big data challenges are not unique to climate science but occur across a wide range of data science applications. Innovative techniques developed by the AI and machine learning community show huge potential but have not yet found their way into climate sciences – and climate scientists are currently not trained to capitalise on these advances. The central hypothesis of IMIRACLI is that merging machine learning and climate science will provide a breakthrough in the exploration of existing datasets, and hence advance our understanding of aerosol-cloud forcing and climate sensitivity. Its innovative training plan will match each ESR with supervisors from climate and data sciences as well as a non-academic advisor and secondment and provide them with state-of-the-art data and climate science training. Partners from the non-academic sector will be closely involved in each of the projects and provide training in a commercial context. This ETN will produce a new generation of climate data scientists, ideally trained for employment in the academic and commercial sectors.