AI Demand Forecasting: Building a Scalable Machine Learning System for Thousands of Products

Explore the development of a scalable AI demand forecasting system at ASOS in this 41-minute conference talk from the Data Science Festival. Dive into the challenges faced by Eduardo Nigri, Priscilla Fearn, and Thomas Buffagni as they built a machine learning system to accurately predict future demand for thousands of products. Learn about the importance of demand forecasting for maintaining healthy stock levels in retail, and how ASOS tackled issues such as sales seasonality, new product launches, and historical data gaps. Gain insights into the journey of creating an AI system now used by hundreds of merchandisers for data-driven decision-making. Discover solutions to modeling challenges, scalable data processing pipelines, MLOps automation, software engineering best practices, and integration into merchandiser workflows. Suitable for those with a basic understanding of machine learning, acquire comprehensive knowledge on implementing and maintaining an effective AI demand forecasting system at scale to deliver tangible business value.