OpenAIBased on real interviews
Checkpoint restore bug hunt
Find bugs that can lose or expose the wrong model-run state during shutdown and restore.
Unlock with Full AccessThe challenge
A class keeps a model run's state in memory and saves it to a checkpoint when it shuts down. Review how it moves between online and offline states, persists updates and restores saved data. Your findings should explain where the implementation violates the supplied lifecycle rules.
What you'll practise
Durable state
Trace whether the latest in-memory changes survive a shutdown.
Restore behaviour
Follow what happens when a checkpoint loads successfully or a restore fails.
Lifecycle results
Check the values returned for offline access, missing keys and stored null values.
Checkpoint code reviewCode included in the full challenge
Your findings
Explain the state-handling bugs and how you would correct them.
The code, instructions and feedback are included with Full Access.