Fashionable, Flexible, and Fun: Examples of Practical AI at Scale
Many will agree that “Knowledge is the currency of 21st century”. Helping turn information into knowledge and thus create value at scale, AI is quickly becoming a fashionable means for organizations across sectors of the economy to assist with decision making and improve products and services. However, with AI an active area of current research and annual cloud compute costs reaching 7-8 digits, discipline is required to leverage its flexibility, and deliver AI-enabled economic results to customers across sectors cost effectively. Here we draw on a few examples from sectors in Canada and beyond, to discuss approaches to manage cost while remaining as scientifically rigorous as possible and delivering measurable results in large scale AI implementations. Using completed large scale projects as examples, we discuss how our teams developed tactics for implementation and repurposing of effective AI algorithms in cloud and on premise big data environments, as well as applied specific popular deep learning research results from recent years – all while having fun along the way.