Dr Torben Sell is leveraging his expertise on geographic data to develop advanced AI techniques to help Aquallite quality assure seafloor datasets for the offshore wind industry. About Aquallite DataAquallite’s subsea data integration software system works alongside geospatial analysts to build highly accurate, engineering-grade models using big data analytics – rather than relying solely on human-judgement. Aquallite’s enterprise software solution will enable collaboration across the supply chain, integrate datasets at scale, and assure quality to generate engineer-grade models. Using machine learning to quality assure bathymetric data2023Aquallite propose an industry MSc project2024PhD student intern funded via innovation voucher scheme2025Aquallite initiate market testing of a micro-service that will eventually be incorporated into its overall product.MotivationThe offshore wind industry is growing at an exceptional rate. However, this has not been without failure – there's been an exponential increase in insurance claims that could put the financial viability of the industry at risk.A major contributing factor is the accuracy of the available 3-dimensional seafloor models – assembled from multiple datasets, each with millions of data points and aligned manually – on which critical planning and engineering-design decisions are based.The quality of this data is assessed manually and, as a result, disparities between the model and the actual seafloor cause significant unplanned costs and delays. Our expertiseTorben Sell is an expert in statistical learning, focussing on understanding the foundations of machine learning and its use in critical applications where a wrong prediction may have a potentially large impact on e.g. people, society, or the economy. His past research includes work on noisy and missing data, Bayesian machine learning, and nonparametric statistical learning methodology, as well as applied research on clustering and classifying satellite images. Our researcher recommendationsThe task of quality assuring seafloor datasets is not suitable for humans, as each dataset includes millions of datapoints and is spread across kilometres of physical distance. Torben is leveraging his expertise on geographic data to develop advanced data analytics techniques, including machine learning, for Aquallite.Why work with us?If you're developing novel technology that requires significant innovation in machine learning, our researchers have access to the skills, systems, and structures you need.Our experts can understand what leading-edge research has been conducted in your area and have access to wider network of people with complementary knowledge and experience. We're keen to partner with early-stage startups like Aquallite, for whom commercial providers of these services would be prohibitively expensive.To discuss any opportunities with us, please email our Business Development team:Email the Business Development team This article was published on 2025-09-04
