1. A real-time endpoint serves a latency-critical feature but the team also needs to periodically score a huge backlog cheaply. What is the BEST combined approach?
- A. Use a real-time endpoint for live requests and batch transform for the periodic bulk backlog✓ Correct
- B. Use one 24/7 real-time endpoint for both
- C. Use batch transform for the live latency-critical requests
- D. Train a new model for each backlog record
Explanation
Match the inference option to each need: a real-time endpoint for low-latency live requests, and batch transform for cost-efficient offline bulk scoring. Forcing one endpoint for both (B) is costly/awkward, batch transform can't meet live latency (C), and per-record training (D) isn't inference.