Databricks OOD Question: Scalable Playlist System Design

You are asked to design a scalable song playlist management system for a music streaming service. The core feature set includes creating playlists, adding/removing songs, and reordering tracks while meeting scalability, performance and consistency goals.

Start by modeling the domain: define User, Playlist, Song, and PlaylistItem (tracks with position/index). Consider storing playlist order with a doubly-linked list plus an index map for O(1) insert/remove, or use fractional/sequence numbering (gap-based ordering) for efficient reordering with O(log n) on index updates. For large playlists, support pagination and cursor-based traversal; apply the Iterator pattern to traverse items safely. Use the Repository pattern to abstract persistence (SQL/NoSQL choices) and the Composite pattern if you support nested playlists or folders.

Interview flow: you’ll be asked to clarify requirements (duplicates allowed? collaborative edits?), sketch data schema and APIs, discuss complexity/performance, then address concurrency and scalability (sharding, caching, optimistic concurrency or CRDTs for collaborative editing). You should propose API endpoints (e.g., POST /playlists/{id}/songs, DELETE /playlists/{id}/songs/{songId}, PUT /playlists/{id}/reorder) with clear request/response shapes and error handling for invalid positions, conflicts, and rate limits.

Skill signals: demonstrate object-oriented design, data modeling, trade-offs between consistency and availability, concurrency control strategies, and practical engineering choices for low-latency operations at scale.