A company is developing a new online application to gather information from customers. An ML engineer has developed a new ML model that will determine a score for each customer. The model will use…

Question

Options

• AConfigure batch transform.
• BConfigure a real-time inference endpoint.
• CConfigure a serverless inference endpoint.
• DConfigure an asynchronous inference endpoint.

• A
  7% (3)
• B
  77% (34)
• C
  5% (2)
• D
  11% (5)

Explanation

Real-time inference (B) is the right choice because the application must score each customer synchronously during their session and immediately display a product - that demands the lowest possible latency with a persistently running endpoint. Batch transform (A) processes large datasets offline in bulk with no live endpoint, making it useless for per-request, interactive scoring. Serverless inference (C) scales to zero between requests and suffers cold-start delays, which directly violates the latency requirement. Asynchronous inference (D) queues requests and returns results later via S3 - ideal for long-running jobs, but completely wrong when the customer is waiting on-screen for a response.

Memory tip: Match the user experience to the inference type - if a human is staring at a loading spinner, only real-time inference is fast enough. Async = fire-and-forget, serverless = cold-start risk, batch = no endpoint at all.

No community discussion yet for this question.