nerdexam
Google

PROFESSIONAL-MACHINE-LEARNING-ENGINEER · Question #310

Your team is experimenting with developing smaller, distilled LLMs for a specific domain. You have performed batch inference on a dataset by using several variations of your distilled LLMs and stored

The correct answer is C. Create a custom Vertex AI Pipelines component that reads the batch inference outputs from Cloud. Vertex AI Pipelines supports custom components that integrate with your existing workflows and allow you to track artifacts such as evaluation metrics, outputs, and metadata. By creating a custom component, you can tailor the evaluation process to your specific needs. This approa

Submitted by daniela_cl· Apr 18, 2026ML pipeline operationalization

Question

Your team is experimenting with developing smaller, distilled LLMs for a specific domain. You have performed batch inference on a dataset by using several variations of your distilled LLMs and stored the batch inference outputs in Cloud Storage. You need to create an evaluation workflow that integrates with your existing Vertex AI pipeline to assess the performance of the LLM versions while also tracking artifacts. What should you do?

Options

  • ADevelop a custom Python component that reads the batch inference outputs from Cloud Storage,
  • BUse a Dataflow component that processes the batch inference outputs from Cloud Storage,
  • CCreate a custom Vertex AI Pipelines component that reads the batch inference outputs from Cloud
  • DUse the Automatic side-by-side (AutoSxS) pipeline component that processes the batch inference

How the community answered

(54 responses)
  • A
    17% (9)
  • B
    7% (4)
  • C
    72% (39)
  • D
    4% (2)

Explanation

Vertex AI Pipelines supports custom components that integrate with your existing workflows and allow you to track artifacts such as evaluation metrics, outputs, and metadata. By creating a custom component, you can tailor the evaluation process to your specific needs. This approach leverages Vertex AI’s built-in tracking and metadata capabilities to maintain an auditable record of model performance, ensuring reproducibility and comparability across LLM variations.

Topics

#Vertex AI Pipelines#LLM Evaluation#Custom Components#Artifact Tracking

Community Discussion

No community discussion yet for this question.

Full PROFESSIONAL-MACHINE-LEARNING-ENGINEER Practice