Meta system design interview questions

Which levels get a design round

Design questions candidates report

• Design the News Feed

  How to design a ranked social news feed for a billion daily users: scope, traffic and memory math for per-user inboxes, the feed API and data model, fan-out of friend posts through an event stream, candidate generation and multi-stage ranking, privacy filtering at read time, seen-state and new-post signals, and the trade-offs you should name.

  Full walkthrough: News Feed Practise on the design board

• Design Instagram / Stories

  A step-by-step design for a photo-sharing app with a home feed: requirements, scale math for half a billion daily users, the upload API with pre-signed URLs, async resizing into several sizes, a CDN for images, hybrid push and pull fan-out for the feed, the celebrity problem, feed caching and cursor pagination, and the trade-offs to name.

  Full walkthrough: Instagram

• Design Messenger / WhatsApp

  A LADDER walkthrough of a mobile chat system: requirements, scale math for 20 billion messages a day and 200 million open connections, the WebSocket and REST API, a message store partitioned by conversation, then the deep dives: routing between chat servers, offline delivery and push, sent-delivered-read receipts, per-conversation ordering, group fan-out, presence, media and end-to-end encryption.

  Full walkthrough: WhatsApp Practise on the design board

• Design live comments

  A LADDER walkthrough of live comments on streaming video: requirements, scale math that shows why naive fan-out to two million viewers is impossible, the API and data model, then the deep dives: per-stream pub-sub with connection servers subscribing once, batching and sampling for hot streams, partitioned writes, moderation and deletes, late joiners and replay, and failure handling.

  Full walkthrough: Live comments

• Design search typeahead

  Design the typeahead box that suggests the top five queries as you type. Requirements and scale (over a hundred thousand requests a second at peak), the suggest API, a trie with top-K cached at every node, an offline pipeline that rebuilds it from query logs, a small real-time layer for trending terms, replication versus sharding, client-side tricks and the trade-offs.

  Full walkthrough: Search autocomplete

• Design a collaborative editor

  A LADDER walkthrough of a collaborative document editor: requirements, scale math for millions of concurrent editors, the API and an operation-log data model, then the hard parts: why concurrent edits conflict, operational transformation versus CRDTs, one session server per document, durability with an op log and snapshots, failover, offline edits, live cursors, version history and permissions.

  Full walkthrough: Google Docs

• Design trending topics / top-K

  How to find the top K most viewed items over the last minute, hour and day at a hundred thousand events a second: why one heap on one machine fails, partitioning counts by item so local top-Ks merge exactly, two-stage aggregation for viral items, sliding windows from tumbling buckets, count-min sketch for bounded memory, a fast stream path checked by an exact batch path, and what "trending" should mean.

  Full walkthrough: Top K (trending)

Also reported

• Design a distributed key-value cache

How to answer any design question: the LADDER method · How the design round is scored by level

Questions people ask