This course explores the foundations of vector databases for storing, organising, and retrieving embedding-based data using ChromaDB. It focuses on the progression from preparing text and generating dense embeddings to understanding vector similarity, vector database architecture, and managing vector data through a persistent local database.

Through structured lessons and practical demonstrations, you will learn how text is cleaned and tokenised, how prepared text is transformed into dense embeddings, and how vectors are compared using cosine similarity, dot product, and Euclidean distance. You will also generate and batch-encode product embeddings with Sentence Transformers, save and reload embeddings using NumPy and pandas, and examine how vector databases overcome the limitations of file-based vector storage. The course progresses from embedding foundations and similarity measurement to practical vector database management with ChromaDB, emphasizing collections, records, metadata, similarity search, filtering, and persistence. Rather than treating a vector database as a simple storage system, it focuses on understanding how embeddings and supporting information are organised, queried, filtered, updated, and retained across application runs. By the end of this course, you will be able to: - Prepare and tokenise product text for embedding generation - Generate, batch-encode, validate, and save dense product embeddings - Compare embeddings using cosine similarity, dot product, and Euclidean distance - Explain vector database architecture and the core elements of vector storage - Create ChromaDB collections and manage embeddings, documents, and metadata - Perform similarity searches and filter retrieved results using metadata - Read, update, delete, persist, and reload vector records with ChromaDB This course is ideal for aspiring AI engineers, machine learning practitioners, developers, data professionals, and learners interested in vector databases, embeddings, and semantic retrieval. A foundational understanding of Python is recommended; prior experience with ChromaDB, vector databases, or advanced retrieval systems is not required. Join us to learn how to move beyond file-based embedding storage, organise vector data with ChromaDB, and build the foundations of persistent, metadata-aware semantic retrieval systems.