Design a Thematic Channel Generator

Problem statement

Design a thematic channel generator that turns an existing media library into continuous, scheduled 24/7 channels built around moods, topics, genres, or franchises — a "Chill Vibes" music channel, a "Nature Docs" documentary stream, a "Comedy Shorts" rotation, or a single-IP channel devoted to one show. Each channel streams a scheduled sequence of assets with no user skipping, smooth transitions or bumpers between items, optional short ad or promo insertion, and a daily or feedback-driven reshuffle of the lineup.

This is not a playlist page inside an on-demand app. A playlist waits for the user to press play on each item. A thematic channel owns the clock: at any wall-clock instant, the channel has exactly one thing playing for every viewer, the schedule is reproducible, and the playout engine must guarantee that the stream never dead-airs, never plays a revoked asset, and never misses an ad break that pays the bills.

Why the problem is distinctive

On-demand systems optimize request-time ranking and per-user session state. Linear channel systems optimize a global timeline. The hard problems shift: schedule construction (which assets, in what order, under which rotation rules), clock-accurate playout (manifests must reflect wall-clock reality to millisecond tolerance), ad insertion at fixed break positions, guide data synchronization, and channel-level analytics like dwell time rather than click-through. Personalization enters differently too: feedback loops adjust the channel for everyone, or split the channel into variants, rather than re-ranking a home screen.

Public evidence shows the category is real and large. Pluto TV, the FAST (free ad-supported streaming TV) pioneer acquired by Viacom in 2019, reported roughly 71.6 million monthly active users in Paramount earnings disclosures around 2022 and operates hundreds of curated and single-IP linear channels. Tubi, owned by Fox, has publicly reported more than 80 million monthly active users and runs themed channels over its AVOD library. The Roku Channel and Samsung TV Plus operate large FAST lineups on the same principle. These are cited company figures for context, not requirements for our fictional system.

The four architectural planes

1. Catalog and theme plane: asset metadata, tags, embeddings, rights windows, and the thematic definitions that group content.
2. Schedule plane: rotation rules, schedule generation, program guide, versioned lineups, and reshuffle workflows.
3. Playout plane: clock-driven manifest generation, segment serving through the CDN, transitions, bumpers, and server-side ad insertion.
4. Feedback and learning plane: dwell, switch, like/skip events, channel popularity analytics, and schedule tuning.

A strong interview answer keeps these planes separate. The schedule plane may degrade while playout keeps serving the last published lineup. Personalization experiments may run without mutating the validated schedule. And the CDN remains the only component that must keep up with viewer-scale concurrency; the control plane only has to be correct.

Public baseline versus design assumptions

Unless tied to a named company citation, every number in this answer is an explicit design assumption stated where used. The working model assumes a mature service with 2,000 live channels, 25 million monthly active viewers, 5 million concurrent viewers at peak, and a library of 50,000 licensable assets. Treat these as sizing inputs, not claims about any real platform.