As the lead ML Engineer for your company, you are responsible for building ML models to digitize scanned customer forms. You have developed a TensorFlow model that converts the scanned images into…

Question

Options

• AUse the batch prediction functionality of AI Platform.
• BCreate a serving pipeline in Compute Engine for prediction.
• CUse Cloud Functions for prediction each time a new data point is ingested.
• DDeploy the model on AI Platform and create a version of it for online inference.

• A
  76% (22)
• B
  14% (4)
• C
  7% (2)
• D
  3% (1)

Explanation

The requirements are: process aggregated data collected at the end of each day, and do so with minimal manual intervention. AI Platform (Vertex AI) batch prediction is designed exactly for this pattern-you submit a batch prediction job that points to a model and an input data source, the platform spins up infrastructure automatically, runs inference across all records, writes results to Cloud Storage or BigQuery, and tears down the infrastructure when done. It requires no persistent serving infrastructure and can be triggered by a Cloud Scheduler job or a Cloud Function, making it nearly hands-free. Option B (Compute Engine serving pipeline) requires manual provisioning, scaling, and management of VMs-high operational overhead. Option C (Cloud Functions per data point) is designed for real-time, event-driven inference on individual records, not for end-of-day batch aggregation. Option D (AI Platform online prediction) deploys a persistent endpoint for real-time serving, which is costly and inappropriate for a once-per-day batch workload.

No community discussion yet for this question.