Design large video upload for a YouTube-like app
Company: Coupang
Role: Software Engineer
Category: System Design
Difficulty: medium
Interview Round: Onsite
## System Design Prompt
Design the **video upload subsystem** for a YouTube-like product that supports **large video files**.
### Core user flow
1. User selects a large video file (hundreds of MB to tens of GB).
2. Upload should be **reliable** (resume after network drops).
3. After upload, the platform processes the video (e.g., transcode) and eventually makes it playable.
### Requirements
- Support **resumable uploads** and unstable mobile networks.
- Efficiently handle large files (chunking, retries, integrity checks).
- Provide upload status to clients (progress, failure reasons).
- Ensure uploaded content is durable and not lost.
- Basic security: authenticate users, authorize uploads, prevent misuse.
### Scale assumptions (state your own if needed)
- Millions of users
- Spiky traffic (new releases)
- Global uploads
### Deliverables
- APIs, data model, and high-level architecture
- Storage strategy for raw uploads and processed outputs
- Processing pipeline (transcoding, thumbnails)
- Failure handling, idempotency, and observability
- Key trade-offs
Overview: This question evaluates a candidate's ability to design scalable, reliable large-file upload subsystems, testing competencies in distributed storage, resumable transfers, data integrity, processing pipelines, security, and observability.