Summary
Full Transcript
GitHub: https://github.com/knodax-labs-demo/aws-data-and-ml-labs/blob/main/32-recommendation-system-with-sagemaker-factorization-machines.md In this hands-on lab, you’ll learn how to build a recommendation system using Amazon SageMaker’s built-in Factorization Machines algorithm—step by step, from scratch. This video walks through the complete end-to-end workflow, starting with creating a SageMaker Domain and Studio environment, generating synthetic user–item interaction data, and preparing the data in the sparse RecordIO-Protobuf format required by SageMaker Factorization Machines. You’ll then train the model using a SageMaker training job, deploy it to a real-time inference endpoint, and generate recommendation scores using live predictions. Along the way, you’ll understand how SageMaker Spaces and JupyterLab work, how IAM roles and S3 permissions affect your workflow, and how to interpret model inference results—especially the difference between prediction scores and predicted labels in recommendation systems. This lab is ideal for: Beginners to Amazon SageMaker Learners exploring recommendation systems AWS Machine Learning exam candidates Anyone wanting hands-on experience with built-in SageMaker algorithms 🔹 What you’ll learn in this video: Setting up SageMaker Studio and Domains Generating synthetic recommendation data Preparing sparse data for Factorization Machines Training and deploying a SageMaker FM model Running real-time inference and ranking recommendations Cleaning up AWS resources to avoid charges ⚠️ Important: SageMaker resources are not free. The video also covers proper cleanup steps to avoid unnecessary billing. If you’re looking for a practical, real-world introduction to recommendation systems on AWS, this lab is a great place to start. 👉 Don’t forget to like, subscribe, and check the description for related labs and resources.
Continue this lesson in the app
Install CourseHive on Android or iOS to keep learning while you move.
