OpenAI System Design Example | Real Interview Question

Problem

Design a video-sharing platform that supports large uploads and on-demand adaptive playback

Accept resumable multipart uploads and ensure idempotent completion without duplicating assets
Asynchronously transcode raw videos into segmented renditions and manifests before making them watchable
Handle extreme read skew on viral videos by offloading segments to a CDN while shielding the origin

Key Test Points

Idempotent upload completion
Staged packaging pipeline
CDN origin shielding

1🎬 What is YouTube?Chapter 1

YouTube is a video-sharing platform where creators upload recorded videos and viewers watch them on demand across many devices and network conditions. From a system-design angle, it combines two very different problems: ingesting very large user uploads reliably, then serving the same content back as low-latency segmented streams at extreme read skew.

2RequirementsChapter 2

3Data model designChapter 3

4API designChapter 4

5High-Level DesignChapter 5

6Deep divesChapter 6

7Other ConsiderationsChapter 7

8Interviewer expectationsChapter 8

Your progressLogin to save progress

Spot an issue or have a suggestion?