grounded in: Krikorian, 'Timelines at Scale', QCon (2012) · Twitter Eng: 'Manhattan' (2014), infra (2017) · twitter/the-algorithm, open source (2023) · Kleppmann, 'Designing Data-Intensive Applications' (2017)
AlgoPlus//structures / case-study
Read the theory

News Feed (Twitter / X)

The home timeline: fan-out-on-write to Redis (2012–17), then the 2023 'For You' ML ranking pipeline.

Step
1/1
Phase
Legend
Design decision
Bottleneck
Solution added
AI Tutor Workspace
In a nutshell
Real systems aren't designed all at once — they grow. You start with the simplest thing that works (one server, one database), watch where it strains under load — the bottleneck — and add exactly one piece to relieve it: a cache, more servers behind a load balancer, a queue to do slow work in the background. Every box on the diagram earns its place by fixing a specific problem, trading a little added complexity for a concrete scaling win.
Ready
Press play to begin the cinematic walkthrough.
Real architectures aren't drawn all at once — they grow. Each component is added to kill one specific bottleneck, trading complexity for a concrete scaling win.
Key terms
Go deeper in the lesson
Read the full theory, intuition & complexity for News Feed (Twitter / X).