Both phone rounds focused closely on Android. The first used a ChatGPT usage-limit scenario: define an API response that lets the client show the affected model, an upgrade link, and retry timing in a Snackbar. Localization mattered, along with how consecutive errors would appear and how Compose should trigger the message.
The second round combined Compose UI with data handling. I needed a rating form with one to five stars, a review field, and a submit button enabled only when both inputs were present. After submission, the interface needed to show confirmation.
These exercises required familiarity with Compose behavior under time pressure, beyond describing the screens. I felt neither round went particularly well and received a rejection soon afterward. The usage-limit details were interview requirements, not a description of live product quotas.
Discussion
Loading comments…