Twitter/X Engineering

Overview

Twitter, now X, is the canonical real-time public timeline product, processing roughly half a million tweets per minute that have to fan out to followers globally within seconds, and is one of the few major platforms that has open-sourced its production recommendation pipeline. The blog's most valuable content lives in three areas: the For You algorithm including candidate generation, heavy ranking, and the SimClusters embedding model, real-time ads infrastructure that serves and paces millions of ad impressions per second with strict budget guarantees, and analytics on Apache Druid serving sub-second aggregations over event streams. For interviews, 'Twitter's Recommendation Algorithm' is a rare end-to-end reference for production recommender design, and 'Storing and retrieving millions of ad impressions per second' anchors high-throughput counter and ledger discussions.

Interview questions

• What topics does Twitter/X engineering blog cover?
• Why is Twitter's recommendation algorithm post important?
• Which Twitter/X posts are most useful for ads-infrastructure interviews?

More Engineering Blogs topics

• Snap Engineering
• Spotify Engineering
• Stripe Engineering
• Swiggy Engineering
• Tinder Engineering
• Twitch Engineering
• Uber Engineering
• Walmart Engineering
• Zomato Engineering