Foo + Bar

In the previous article, we learned how we can test for similarity between two pieces of data through vectorizing and comparing vectors. However, we left off with a linear search to test for vector similarity. This does not scale since our workload increases linearly as the amount of content gets added.

In this article, we will use FAISS (Facebook AI Similarity Search) to optimize our vector search. FAISS was built by the legends at Meta and was developed to be capable of handling large scale vector searches.

Have you ever found yourself scrolling through Instagram or Netflix and thought to yourself, jee, I wonder how it knows what I like and what to recommend? In this article, I aim to provide a shallow dive into content recommendation systems and share the method I’m using to build my current startup idea.

At the end of this article is a functional demo where you can see a very simple implementation of a recommendation system.