Mock Interviewer
Ask Coach
Tap a question to start, or type your own below.
YouTube System Design
Visual Problem Diagram
Scenario
A 4K upload finishes encoding 20 minutes after the user shared the link—viewers hit 404 segments until manifests update. Trending page loads must not scan every view event in real time. Upload pipeline, CDN playback, and view counting are separate scaling problems the interview expects you to separate.
Design a user-generated video platform with upload, transcoding, search, playback, and view statistics. UGC implies spiky uploads and long-tail storage; playback reuses CDN + adaptive streaming patterns.
You should support chunked upload, async encode, catalog metadata, search, smooth playback, and scalable view counts. Be ready to separate write (upload) from read (play) from aggregate (views) paths.
Constraints
Upload video, transcode, store metadata, search, stream playback, view/like counts, comments hook
< 3 s playback start p95, durable uploads, 99.9% availability, approximate view counts OK
2B users, 500 hours uploaded/min, billions of views/day, petabyte storage
Stages ahead
Log in to start practicing
Preview the problem below. Sign in to start your timed session, save progress, and use the AI coach.