What is Interview Forge?

Interview Forge helps you practice mock interviews, follow study plans, track career goals, review learning analytics, and prepare across 36 pages.

Design a News Feed (Social Feed)

Requirements & scope

Functional: users post content; followers see a ranked feed of recent posts; like/comment.
Non-functional: feed loads in <200ms, new posts appear within seconds, ranking is pluggable.
Scale: ~1B users, ~100M posts/day, read-heavy.
Consistency: feed ordering can be eventually consistent; a post may take a second to appear.

Out of scope (covered in detail below)

Ads injection, the ranking model training pipeline, content moderation, and A/B testing infra — deferred and discussed in the dedicated section.

Capacity estimation

Resource	Estimate	Notes
Posts/day	~100M	Each a small object
Feed reads/day	~billions	Read-heavy, ~100x writes
Feed cache (Redis)	~terabytes	Pre-computed feed lists
Max followers	~tens of millions	Celebrity problem

Architecture overview

News Feed — fan-out -> cache -> read

Hover a node to highlight its connections, or click a node to see its role. The fan-out worker pre-computes feeds; reads are a cached list fetch + batch object load.

ClientServiceDatastoreCacheQueueExternal

The core decision: fan-out vs pull

Fan-out on write: on each post, push to every follower's pre-computed feed. Fast reads; expensive for celebrities.
Pull on read: fetch recent posts from followed users and rank on the fly. Cheap writes; heavier reads.
Hybrid: fan-out for normal users, pull for celebrities — the production answer at scale.

Unlock the full guide

Premium members get every DSA, system design, company and role interview guide — plus AI practice, mock interviews, and analytics.

Unlock with Premium