A buttery hatch already beat its word-error target in Domain 1. Meal hour now triples turns. What should the candidate do?
Select an answer to reveal the explanation.
Short Explanation
Quality homework is done; meal hour is a crowd problem. Scale the chart for the rush—do not redo the lab notebook when the scores already passed.
Full Explanation
A buttery hatch that already beat its word-error target in Domain 1 and now faces meal hour tripling turns should use the Helm chart to scale for load rather than reopening the experimental vocabulary notebook. Quality wins already landed; capacity is the meal-hour lever.
Reopening the Domain 1 vocabulary notebook as the first meal-hour response fails when WER already meets target and load is the new variable. Ignoring load and lowering quality targets until the line disappears fails public service. Inventing a NIM serving tier because word-error wins require a new product name fails by inventing brands unrelated to the scale need.
Exam caveat: after Domain 1 quality targets are met, meal-hour lines are Domain 6 scale. Do not regress into vocabulary notebooks or invented serving tiers first.
Operational check: lock WER dashboard as still-green; scale Helm replicas for meal hour; re-measure turns and delay at the municipal buttery; leave vocabulary experiments closed unless quality regresses. In municipal platform operations, record the chosen performance lever and the rejected alternatives in the change record so future on-call engineers understand the failure modes avoided for citizen services. Re-check the decision after the next traffic spike or model update to ensure the same answer still holds for the city's estate.