Abstract: Deploying models in to production is still the hardest problem in Data Science, but not if you use the right tools. In this session we look at how MLFlow pipelines simplify the process of deploying models to production. We look at the challenges of data management and how Delta can be used to ensure model are reproducible, without taking many replicas of your dataset.
Bio: Terry is the Director of AI for Advancing Analytics and Microsoft Artificial Intelligence MVP with a focus on all things AI and Data Science. Terry has a passion for applying traditional Software Engineering techniques to Data, to improve the way teams deliver Machine Learning projects. Terry is the host of the popular podcasts Data Science in Production and Totally Skewed, and organises the Global AI Bootcamp London event.