Vector Databases for Beginners
Learn what a vector database does, when you need one and how to choose without falling for marketing.
A taste of a lesson
We have about 20,000 help articles split into chunks. Do we need a vector database?
Probably not yet. Twenty thousand articles might become around a hundred thousand chunks; with a few hundred dimensions each that fits in memory, and exact search with one matrix multiplication is usually quick. If you already run a relational database, its vector extension could store vectors next to your articles with filters and backups you know. Move to an approximate index or dedicated system when measurements show you need it. How often do your articles change, and do users filter by product or language?
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- Explain what nearest neighbour search and vector storage do
- Build a small brute force vector search with plain arrays
- Describe how approximate indexes trade recall for speed
- Compare dedicated, extension, search engine and in process options
- Handle filters, updates and deletes correctly in a vector index
Lesson plan
- 1 What a vector store does Describe storing vectors with metadata and finding nearest neighbours. Start
- 2 Brute force first Build exact vector search with arrays and measure it. Start
- 3 Approximate indexes Understand how ANN indexes speed up search at some cost in recall. Start
- 4 The landscape of options Compare categories of vector storage tools on practical criteria. Start
- 5 Filters, updates and deletes Keep results correct as data changes and users filter. Start
- 6 Making the choice Decide on a setup for your project and justify it. Start
Try asking
About this tutor
For beginners building semantic search or RAG who keep hearing that they need a vector database. You learn what vector storage and nearest neighbour search actually do, why plain arrays are enough for small collections, how approximate indexes trade a little accuracy for speed, and what the real options look like: dedicated vector databases, vector support added to databases you may already run, search engines with vector features and in process libraries. You compare them on scale, filtering, updates, hybrid search and operational effort, and you build a small working index yourself before choosing.
Reviews
4.3
3 ratingsSample
- Georg H.Sample
Building brute force search in a few lines showed we did not need a new database for our size. Saved us weeks of setup.
- Lerato M.Sample
Clear comparison of categories without pushing any vendor. The filter during versus after search point fixed our missing results.
- Pierre D.Sample
Good beginner pace. I would have liked a little more on quantisation, but the essentials are all here.
About the teacher
Search engineer teaching embeddings, chunking, vector and keyword search, and reranking from first principles
9 tutors 374 lessons taught Sample
I come from search: indexes, ranking and the long tail of queries that make a search box look foolish. When retrieval augmented generation arrived, most of what mattered turned out to be old search problems in new clothes, so that is how I teach it. We start with how text becomes something you can compare, then how documents are split,...
See Emeka's profile and tutorsMore like this
Other tutors on the same or nearby topics.