Design a Caching Strategy

Background: Native [platform] app. This data is expensive or slow to
fetch: [describe—e.g., feed, images, reference data]. How fresh it needs
to be: [real-time / minutes / hours / rarely changes].

Use case: Design a caching strategy that makes the app feel instant
without serving dangerously stale data.

Implementation details:
- Layered cache: in-memory (fast, volatile) over on-disk (survives
  relaunch); decide what belongs in each
- A freshness policy: TTL, ETag/Last-Modified conditional requests, or
  cache-then-revalidate (stale-while-revalidate)
- Invalidation: how and when entries are evicted or refreshed, including
  on user-driven actions (pull to refresh, logout clears user data)
- iOS: where URLCache, NSCache, and on-disk store each fit. Android:
  OkHttp cache, in-memory cache, and DataStore/Room for structured data.
- Image caching specifically, if relevant (sizes, memory pressure)

Layer & dependencies: Caching is a data-layer concern hidden behind the
repository; callers just ask for data.

Deliverable: A caching design with the layers, the freshness/
invalidation policy per data type, the eviction rules, and the
memory-pressure handling.