We're selecting the next generation of engineers who govern AI, not just use it. Apply now for the only certification of its kind.
Answer these 4 scenarios honestly — there are no right or wrong answers. We want to see how you think, not what you know.
Take your time. You can write in any language you're comfortable with.
An AI agent flags a 12% confidence drop on a production task it's been handling reliably for weeks. The metrics dashboard shows no other anomalies. What's your first move, and why?
A client reports their AI agent 'feels wrong' — outputs are technically correct but the client is uneasy. Metrics look normal. How do you investigate this?
You notice a pattern: every Tuesday at 3pm, your AI agent's quality scores dip for about 90 minutes, then recover. This has happened 4 weeks in a row. What's your hypothesis, and what would you check?
You're paired with a senior engineer who disagrees with your monitoring approach. They want to increase the confidence threshold; you think the issue is upstream data quality. How do you resolve this?