How the interviewer opens it
“Design a platform like Pinterest where users save images to boards and discover related content. The service must serve image-heavy pages efficiently at large scale.”
The prompt is vague on purpose. The first minutes are for you. Reduce the prompt to the core features and agree on the numbers before you draw.
What to clarify first
- Product behavior
- Scale and request mix
- Traffic distribution
- Image sizes
- Latency
- Related content
- Board access
- Consistency and availability
- Durability and deletion
- Exclusions
The interviewer has the answers. In a SysRound round, Alex gives these answers, as a real interviewer does.
Scope of a strong answer
Core requirements
- Create a pin with an image and a source link.
- Save pins to public or private boards.
- View pins and board contents.
- Discover related public pins.
Usually out of scope
- Web crawling
- Personalized home feeds
- Image model development
- Ads and chat
- Shared boards with several editors
What interviewers ask about
- Images on a page
- Image bytes dominate traffic. New uploads also need several steps before the service can display them safely.
- Saves and board access
- Many boards can reference one pin. Save retries, concurrent edits, and private media require distinct rules for ownership and access.
- Related content
- The service must find useful matches across billions of images. New uploads lack user activity, and stale results can expose unavailable content.
Common mistakes
- The design sends original images for every thumbnail.
- The design copies an image each time a user saves its pin.
- The design publishes a new pin before its image variants exist.
- The design uses large offsets for board pages.
- The design scans the full image collection for each related-content request.
- The design treats a secret image URL as sufficient protection for private content.
The bar by level
- Senior (L5/E5)
- Quantifies image traffic and storage. Defines media states, retry behavior, cursor pagination, and shared pin deletion. Builds bounded discovery requests with an asynchronous index.
- Staff (L6+/E6+)
- Defines access guarantees across APIs, indexes, and CDN caches. Explains partition growth, regional failure, and recovery. Connects cache hit rates and image variants to cost.