Drift
A hard AI Coding mock interview question on DataDriven, with an AI interviewer who asks follow-up questions and a hire or no-hire decision at the end.
- Domain
- AI Coding
- Difficulty
- hard
- Seniority
- L6
Interview Prompt
Each day the history store gains one summary of the same pinned probe set, embedded through a hosted endpoint whose vendor can swap the model behind an unchanged URL, and tests/test_drift_alarm.py replays one such swap as an 18% jump in cosine_to_anchor_mean. detect_drift() in monitoring/drift.py returns whether today's summary alarms and a reason naming the metric that moved, holding to the scored metrics, baseline length, threshold and false-positive budget fixed in the read-only monitoring/constraints.sql. A swap fires in whichever scored metric it moves, even one that has sat perfectly still, but a day that stays inside the ordinary day-to-day noise of its recent baseline is quiet, however far it lands from yesterday.
Summary
Nothing changed. Everything moved.
How This Interview Works
- Read the task and the repository: the files, the failing check and the constraints
- Work the fix with the AI agent at your side, and run the checks until they pass
- Submit, then answer the interviewer's questions about the change you made
- Get a hire or no-hire decision with feedback on how you drove the agent