Free Databricks Databricks-Certified-Generative-AI-Engineer-Associate Exam Questions

Become Databricks Certified with updated Databricks-Certified-Generative-AI-Engineer-Associate exam questions and correct answers

Page: 1 / 30

Total 150 Questions | Updated On: Sep 13, 2024

Add To Cart

Question 1

As a Gen AI engineer, you are preparing to evaluate your language model both before and after deployment. Which of the following approaches best describes the key focus of online evaluation compared to offline evaluation?

A : Offline evaluation focuses on analyzing real user behavior data collected from live deployments, while online evaluation involves curating benchmark data and using task-specific metrics.

B : Offline evaluation is concerned with collecting user feedback through live interactions, whereas online evaluation relies on predefined benchmark datasets and task-specific metrics.

C : Online evaluation requires extensive preparation of benchmark datasets before deployment, while offline evaluation is centered on real-time monitoring of user interactions with the model.

D : Online evaluation involves deploying the app and assessing the model based on real user interactions and feedback, whereas offline evaluation is conducted using curated benchmark data and task-specific metrics.

Answer: D

Question 2

You are developing an AI application that needs to handle and generate diverse types of content for an interactive storytelling platform. The platform must integrate text and images seamlessly, allowing users to generate stories with accompanying visuals. Which of the following approaches best leverages Multi-Model Architectures for this scenario?

A : Embed text and images into the same vector space but do not use any advanced Multi-Modal LLMs, relying on basic retrieval systems to handle each modality individually.

B : Use one modality, such as text, as the primary focus and adapt images according to predefined templates or rules, without dynamically generating or integrating images with the text.

C : Use a Multi-Modal LLM like GPT-4V(ision) to generate both the narrative text and associated images in a unified manner, allowing the system to produce a story with integrated visuals based on user input.

D : Store text and images separately in distinct databases, and use separate models for text generation and image retrieval, without integrating them into a single system.

Answer: C

Question 3

How can a Generative AI Engineer handle complex documents (with mixed text and images) for data extraction and chunking for a RAG based LLM application?

A : Using python libraries like PyPDF

B : Using Multi-modal models like OpenAI's GPT-4o and beyond

C : All of these options

D : Using layout model from Hugging Face

Answer: C

Question 4

Which of the following parameters can affect the determinism and creativity of an LLM?

A : All of these options

B : "Top-k" value

C : "Top-p" value

D : "Temperature" value

Answer: A

Question 5

A Generative Al Engineer has already trained an LLM on Databricks and it is now ready to be deployed. Which of the following steps correctly outlines the easiest process for deploying a model on Databricks?

A : Log the model as a pickle object, upload the object to Unity Catalog Volume, register it to Unity Catalog using MLflow, and start a serving endpoint

B : Log the model using MLflow during training, directly register the model to Unity Catalog using the MLflow API, and start a serving endpoint

C : Save the model along with its dependencies in a local directory, build the Docker image, and run the Docker container

D : Wrap the LLMs prediction function into a Flask application and serve using Gunicorn

Answer: B

Page: 1 / 30

Total 150 Questions | Updated On: Sep 13, 2024

Add To Cart