Challenge · intermediate · 25 min · Premium
Structured output, or nothing
Calder's agent sometimes answers in prose where a proposal object is expected, and the pipeline counts those as successes with an empty action. Count them, make the output a schema or a hold, and stop success from meaning nothing happened.
Brief
- Customer
- Calder Equipment (synthetic customer from the Production AI Agents course; every sample and log on this box is made up for the exercise)
- Situation
- Last night's run recorded 40 successes. Operators found that several of those cases have no proposal at all: the model wrote a paragraph instead of the proposal object, the parser returned an empty action, and the pipeline logged success because nothing threw. Nobody was told.
- Your role
- You are the FDE who owns the output contract. You have the recorded outputs, the validator config (bound to the live runtime), and a validator you can run over the samples.
- Goal
- Count the outputs that carry no proposal, make the runtime require the schema and hold on an invalid output, and show the validator reporting held cases instead of empty successes.
Objectives
- 1.Count the outputs that are not a proposal (you commit to an answer)
- 2.Require the schema and hold on an invalid output
- 3.Show the validator reporting held cases and zero empty successes
- 4.Say why success without a payload is a failure (you commit to an answer)
Mission workspace
This mission is part of premium, with every challenge and incident in the lab. The brief above is the whole problem; the box, the hints and the debrief are what you unlock.
Get full Premium access · ₹2,000 / $25
Every answer, concept and course, plus Premium PDF guides, companion files, the full practice-test bank and work-sample downloads. Referral Premium excludes guide PDFs and their companion files.
6 months · One payment · No auto-renewal
Study alongside free video lessons.