A company has an experimental ML model that must be validated on production traffic without impacting the current live model. Only one model can handle user requests at a time. Which deployment approach lets you evaluate the new model’s predictions in production without affecting live traffic?

Shadow deployment is correct because it routes a copy of live production traffic to the new model without the new model's predictions impacting the actual user responses. This allows for real-time validation and performance monitoring of the experimental model using production data, without risking service disruption or incorrect results for users. A/B testing would route a portion of live traffic to the new model, directly impacting user experience. Canary release gradually rolls out the new model to a small subset of users, also impacting live traffic. Blue/green deployment switches all traffic from the old version to the new version at once, which is too risky for an experimental model that cannot impact live traffic.