We’re excited to announce the launch of our new feed algorithm, codenamed Sauron.

What is Project Sauron?

Project Sauron introduces a cutting-edge Two Tower retrieval model, a technique famously used by platforms like YouTube, Instagram, and Uber Eats. This state-of-the-art approach enables lightning-fast personalized retrieval of content from a database containing millions of records. In our context, this model helps us connect you with the best development content available.

How Does It Work?

The Two Tower model creates a new embedding space where both our posts and users coexist. This space allows us to calculate similarities between entities—posts and users—enabling us to recommend content and users that align with your interests. The magic lies in generating these embeddings offline, so in real-time, we only perform vector similarity computations using Redis. This process ensures incredibly fast feed delivery, happening within microseconds.