Types of inference · Question 11 of 12
An insurer offers an internal review tool that is used only a few times a day. Each request is small, an answer after one or two seconds is fine, and costs should be minimal. Which option is most suitable?
1 answer is correct. Hard
Show answer
Answer
- A Serverless inference
Explanation
With rare, small requests a permanently running endpoint would sit idle most of the time and still incur costs. Serverless inference scales to zero when idle and bills only the actual processing time. Occasional cold starts are acceptable here because a one-second delay is tolerated.