Databricks Certification Practice Test 11

Practice Test 11

Question 1 of 30

📋 Multiple Choice

Question 231

## Question 231 A retail company has deployed a Large Language Model (LLM) chatbot for customer support. Customers report that the chatbot occasionally provides incorrect return policies. What is the MOST appropriate recommendation?

A. Increase cluster size

✓

B. Add Retrieval-Augmented Generation (RAG)

C. Create more workspaces

D. Reduce notebook count

Answer

✅ B. Add Retrieval-Augmented Generation (RAG)

Explanation

RAG allows the chatbot to retrieve current company policies before generating answers. ---

Answered: 0 / 30

Databricks Certification Practice Test 11

Practice Test 11

Insightful Saga — Modern Data Engineering Certification Preparation

Question 231

A. Increase cluster size

B. Add Retrieval-Augmented Generation (RAG)

C. Create more workspaces

D. Reduce notebook count

Answer: ✅ B. Add Retrieval-Augmented Generation (RAG)

Explanation:

RAG allows the chatbot to retrieve current company policies before generating answers. ---

Question 232

## Question 232 A banking organization wants an AI assistant to answer questions using internal documents without retraining the model every week. What is the BEST approach?

A. Retrain daily

B. Store documents in Excel

C. Implement Vector Search

D. Create more catalogs

Answer: ✅ C. Implement Vector Search ---

Question 233

## Question 233 A healthcare provider needs to identify whether an LLM-generated answer originated from approved company knowledge sources. What capability should be prioritized?

A. Response Traceability

B. More GPUs

C. More Partitions

D. Dashboard Scheduling

Answer: ✅ A. Response Traceability ---

Question 234

## Question 234 A company wants AI-generated responses to cite supporting documents. What is the BEST recommendation?

A. Use manual screenshots

B. Enable source attribution

C. Increase worker nodes

D. Disable monitoring

Answer: ✅ B. Enable source attribution ---

Question 235

## Question 235 A customer service chatbot sometimes generates fabricated information. This problem is commonly known as:

A. Serialization

B. Lineage

C. Hallucination

D. Partition Drift

Answer: ✅ C. Hallucination ---

Question 236

## Question 236 An enterprise wants to prevent employees from sending confidential information to external AI systems. What should be implemented?

A. AI Governance Controls

B. More Storage

C. Additional Dashboards

D. More Pipelines

Answer: ✅ A. AI Governance Controls ---

Question 237

## Question 237 A legal department requires every model deployment to pass compliance reviews before production release. What is the BEST approach?

A. Manual Emails

B. Governance Approval Workflow

C. Larger Clusters

D. More Repositories

Answer: ✅ B. Governance Approval Workflow ---

Question 238

## Question 238 A logistics company wants to measure whether generated AI responses are helpful. What should be collected?

A. User Feedback Metrics

B. Cluster Names

C. Table Comments

D. CPU Count

Answer: ✅ A. User Feedback Metrics ---

Question 239

## Question 239 A retail recommendation engine stores product descriptions as embeddings. What is the PRIMARY purpose?

A. Dashboard Creation

B. Semantic Similarity Search

C. Workflow Scheduling

D. Cluster Monitoring

Answer: ✅ B. Semantic Similarity Search ---

Question 240

## Question 240 An LLM solution must answer only from approved company documentation. What architecture is most appropriate?

A. Open Internet Search Only

B. RAG Architecture

C. Manual Text Entry

D. Weekly Retraining

Answer: ✅ B. RAG Architecture ---

Question 241

## Question 241 A company discovers that a recently deployed AI model responds differently to the same question after multiple releases. What should be investigated?

A. Model Version Changes

B. Column Aliases

C. Dashboard Layout

D. Storage Accounts

Answer: ✅ A. Model Version Changes ---

Question 242

## Question 242 A financial institution wants complete visibility into prompts sent to AI systems. What should be captured?

A. Prompt Logs

B. Cluster Policies

C. Job Names

D. Partition Metadata

Answer: ✅ A. Prompt Logs ---

Question 243

## Question 243 A company wants faster AI response retrieval from millions of documents. What should be introduced?

A. Vector Index

B. CSV Files

C. Static Dashboards

D. More Schemas

Answer: ✅ A. Vector Index ---

Question 244

## Question 244 A support chatbot needs access to recently updated product documentation every hour. What is the BEST recommendation?

A. Monthly Retraining

B. Continuous Knowledge Refresh

C. Delete Old Documents

D. Disable Indexing

Answer: ✅ B. Continuous Knowledge Refresh ---

Question 245

## Question 245 An enterprise AI team wants to compare answer quality across different LLMs. What should be established?

A. Evaluation Framework

B. Additional Catalogs

C. New Visual Theme

D. More Worker Nodes

Answer: ✅ A. Evaluation Framework ---

Question 246

## Question 246 A company wants to identify questions where an AI assistant performs poorly. What should engineers analyze?

A. Failed Interaction Patterns

B. Notebook Colors

C. Cluster Ownership

D. Driver Memory Only

Answer: ✅ A. Failed Interaction Patterns ---

Question 247

## Question 247 An insurance provider requires AI responses to remain consistent with regulatory guidelines. What should be introduced?

A. Guardrails

B. More Clusters

C. More Dashboards

D. More Catalogs

Answer: ✅ A. Guardrails ---

Question 248

## Question 248 A retail company wants to understand which documents contributed to a generated answer. What capability provides this?

A. Retrieval Metadata

B. Checkpointing

C. Repartitioning

D. Compaction

Answer: ✅ A. Retrieval Metadata ---

Question 249

## Question 249 An AI application serves thousands of users simultaneously. Response latency increases significantly during peak hours. What should be reviewed first?

A. Model Serving Metrics

B. Schema Definitions

C. Catalog Structure

D. Dataset Names

Answer: ✅ A. Model Serving Metrics ---

Question 250

## Question 250 A company has multiple AI assistants serving different departments. Management wants consistent governance, monitoring and deployment standards across all assistants. What is the MOST appropriate strategy?

A. Centralized AI Platform Governance

B. Independent Department Standards

C. Separate Monitoring Per User

D. Manual Tracking Documents

Answer: ✅ A. Centralized AI Platform Governance

Explanation:

Centralized governance improves consistency, compliance, security, monitoring and operational control across enterprise AI solutions. ---

Question 251

## Question 251 An airline company plans to build an AI assistant that helps maintenance engineers troubleshoot aircraft issues using maintenance manuals, repair histories, and engineering bulletins. What should be created first?

A. Vectorized Knowledge Base

B. Additional Dashboards

C. New Workspaces

D. SQL Warehouse

Answer: ✅ A. Vectorized Knowledge Base ---

Question 252

## Question 252 A company wants an AI assistant to answer only questions related to HR policies and refuse unrelated topics. What is the BEST recommendation?

A. Domain Restriction Rules

B. More GPUs

C. More Clusters

D. Additional Repositories

Answer: ✅ A. Domain Restriction Rules ---

Question 253

## Question 253 An ML team wants to understand why users are abandoning an AI application. What should be analyzed?

A. Conversation Analytics

B. Cluster Tags

C. Catalog Comments

D. Data Compression Ratios

Answer: ✅ A. Conversation Analytics ---

Question 254

## Question 254 A company deploys a document-search AI solution. Users complain that answers are outdated. What is the MOST likely recommendation?

A. Refresh Source Indexes More Frequently

B. Add More Dashboards

C. Increase Notebook Size

D. Add More Schemas

Answer: ✅ A. Refresh Source Indexes More Frequently ---

Question 255

## Question 255 A GenAI application serves five countries. Management wants to compare response quality by region. What should be implemented?

A. Regional Evaluation Metrics

B. More Storage Accounts

C. More Catalogs

D. Cluster Renaming

Answer: ✅ A. Regional Evaluation Metrics ---

Question 256

## Question 256 A company wants AI-generated content reviewed before being sent to customers. What process should be introduced?

A. Human-in-the-Loop Validation

B. Daily Cluster Restarts

C. Dashboard Archiving

D. Workspace Duplication

Answer: ✅ A. Human-in-the-Loop Validation ---

Question 257

## Question 257 A support organization wants AI systems to automatically escalate uncertain responses to human agents. What capability should be implemented?

A. Confidence Threshold Routing

B. Additional Partitions

C. Bulk Retraining

D. Dashboard Monitoring

Answer: ✅ A. Confidence Threshold Routing ---

Question 258

## Question 258 A company wants to prevent prompt injection attacks against an enterprise AI assistant. What should be prioritized?

A. Input Validation and Prompt Security Controls

B. More Compute Nodes

C. Additional Databases

D. More BI Reports

Answer: ✅ A. Input Validation and Prompt Security Controls ---

Question 259

## Question 259 An enterprise wants to compare the cost and quality of multiple foundation models before selecting one. What should be performed?

A. Benchmark Evaluation

B. Notebook Cleanup

C. Schema Optimization

D. Workspace Migration

Answer: ✅ A. Benchmark Evaluation ---

Question 260

## Question 260 A newly hired Data Engineer joins a company building a Lakehouse AI platform. Management wants a scalable solution where Data Engineering, Analytics, Machine Learning, and Generative AI share the same trusted data foundation. What architectural principle should guide the implementation?

A. Unified Lakehouse Architecture

B. Separate Data Platforms per Team

C. Department-Owned CSV Files

D. Independent AI Data Sources

Answer: ✅ A. Unified Lakehouse Architecture

Explanation:

A Unified Lakehouse Architecture reduces silos, improves governance, promotes reuse, and supports analytics, ML, and GenAI workloads from a single trusted platform. --- --- title: Databricks Certification Practice Test 12 description: Advanced AI & ML Engineering Scenarios ---