(As an extension of the basic application of big data, the recommendation system enterprise engineering practice project is built on the user portrait project. The purpose is to let students learn the basic idea of ??building an enterprise-level recommendation system, and explain in depth the two most important links in the recommendation system, recall and sorting. Each link introduces related algorithms based on Spark-Mllib, such as the recall layer ItemCF, ALS Two-way recall algorithm, the fusion sorting layer introduces GBDT+LR. While understanding the algorithm, pay more attention to engineering practice. We will extract features from raw data. Explain in depth the transformation from algorithm model design to programming implementation, and give practical cases for the cross-platform deployment of algorithm models, so that students can understand the deployment and use of algorithm models in actual projects.)