Microsoft Recommenders: Best Practices for Production-Ready Recommendation Systems
Andreas A. Argyriou, Miguel González-Fierro, Le Zhang
Abstract
Recommendation algorithms have been widely applied in various contemporary business areas, however the process of implementing them in production systems is complex and has to address significant challenges. We present Microsoft Recommenders, an open-source Github repository for helping researchers, developers and non-experts in general to prototype, experiment with and bring to production both classic and state-of-the-art recommendation algorithms. A focus of this repository is on best practices in development of recommendation systems. We have also incorporated learnings from our experience with recommendation systems in production, in order to enhance ease of use; speed of implementation and deployment; scalability and performance.