ARISE: An Artificial Intelligence Studio for Scientists and Engineers

Science has become a data-driven effort: researchers from psychology, engineering, and space science gather and analyse vast amounts of data, using tools that were previously reserved for Computer Science professionals. This is clearly beneficial: scientists from diverse disciplines are learning computing skills that are in high demand while advancing the science of their respective fields. However, there is a huge gap between doing data analysis in experimental setting and deploying data pipelines in a production environment. Besides knowing statistics, and data analysis, deploying systems in the real world requires following rigorous software engineering principles.
Scientists using data analysis are already in a good position for becoming machine learning practitioners: they have good knowledge of the context, a problem, and often sound foundations in mathematics and statistics. Engineering machine learning systems is a complex topic that demand considerable effort to master, but the benefits of incorporating this skill into our academic community are immense.
Within the School of Computing, we have several years of research and industry experience in the engineering of machine learning systems. We envision a training unit (a 'studio') within the school to provide research scientists with the skills for engineering production-ready machine learning systems. Such unit would deliver a curriculum tailored for research scientists, using real use cases gathered from the researchers and scientists themselves.