Learn to build an AI recommendation engine from scratch using real Yelp review data in this hands-on session that transforms flat tabular data into a queryable graph-based system. Discover how to identify hidden entities and relationships within your existing datasets without complex machine learning pipelines, then leverage AWS services including Bedrock, OpenSearch, and Neptune to create a working recommendation system. Master the process of converting rows and columns into meaningful graph structures, implement natural language querying capabilities, and understand when graph traversal approaches outperform traditional ML similarity models for recommendation problems. Gain practical skills in spotting graph opportunities in tabular data, executing data transformations without complex ETL processes, and developing intuition for choosing between graph-based and similarity-based solutions for recommendation challenges.