A curated collection of demos, blog posts, official documentation, and training resources mapped to each exam objective for the Databricks Certified Generative AI Engineer Associate certification (March 2026 version).
How to Use This Guide
For each exam section and objective, this guide provides:
- 📚 Official Documentation: Direct links to official Databricks docs (docs.databricks.com)
- 🎯 Demos: Interactive demonstrations and tutorials
- ✍️ Blog Posts: Technical articles and best practices
- 🎓 Training Resources: Courses, certifications, and learning materials
Resources are ranked by relevance score based on keyword matching. Review multiple resources for each objective to get comprehensive coverage.
About the Author
I’m a Databricks Solutions Architect Champion with extensive experience in AI/ML and generative AI applications. This guide is designed to help you navigate the Generative AI Engineer Associate certification, which focuses on building production-ready GenAI applications on the Databricks platform. I took this exam back in 2024 and am preparing to renew in the coming months, creating this guide for the new syllabus is part of my preparation process.
The GenAI Associate exam tests your practical knowledge of building and deploying LLM-powered applications. As of the March 2026 update, this spans prompt engineering, RAG, Vector Search, MLflow 3 (Tracing, Scoring, Prompt Registry), the Mosaic AI Agent Framework, Agent Bricks, MCP servers, the AI Gateway, and building agent UIs on Databricks Apps.
I created this guide by analyzing the exam objectives and mapping them to the best available resources. My advice: get hands-on. Application Development is the largest section at 30% and Assembling & Deploying is another 22%, so build an agent end-to-end at least once. Pick the Agent Framework or Agent Bricks, wire RAG and Vector Search into it, log with MLflow, deploy on Model Serving or Databricks Apps, and wrap it in AI Gateway. The pieces lock into place when you build the whole thing.
Find out what works best for you. Good luck on your Databricks certification journey!
🆕 What Changed in March 2026
If you sat the April 2025 version of this exam, plan to re-prep. Databricks published a new exam guide on March 18, 2026 with real changes to scope. Section weights are now published. Application Development is the heaviest at 30%, Assembling & Deploying follows at 22%, and Governance is the lightest at 8%, so plan your time accordingly.
Content changes: a new objective on Agent Bricks (Knowledge Assistant, Multi-Agent Supervisor, Information Extraction) joined Section 1. Six new objectives in Section 4 cover MCP servers, prompt version control via the MLflow Prompt Registry, CI/CD for agents, persistent agent memory, storage-optimized Vector Search configuration, and building user-facing agents with Databricks Apps. Section 6 grew to ten objectives, adding AI Gateway tracking (inference and usage tables, rate limiting), custom Scorers via mlflow.genai.evaluate(), and incorporating SME feedback. Older content on prompt-format craft, metaprompts, and standalone “code a chain in LangChain” objectives has been folded into broader items.
MLflow 3 is now central. Its Tracing, Scoring, and Prompt Registry are all newly testable, and the older Agent Evaluation framework is positioned as legacy. The exam grew from 53 to 56 scored objectives, with more weight on production agent operations.
📖 Background Reading
Before diving into the objectives, these resources provide essential foundational context for the Generative AI Engineer Associate exam:
RAG Fundamentals
What is Retrieval Augmented Generation (RAG)? – Read the Blog Post Start here! This foundational post explains how RAG combines retrieval with generation to create more accurate and grounded LLM applications. Understanding RAG architecture is essential for most exam sections.
Building RAG Applications on Databricks – Read the Documentation The core documentation for understanding how to build RAG applications – covers chunking, embedding, retrieval, and generation.
Vector Search & Embeddings
Production-Quality RAG Applications with Databricks – Read the Blog Post A practical look at how Vector Search, embeddings, and retrieval combine in production GenAI apps. Useful background for both Section 2 (Data Preparation) and Section 4 (Deploying RAG).
Vector Search Documentation – Read the Documentation Learn how to create indexes, manage embeddings, and query vector databases.
LLM Application Development
Mosaic AI Agent Framework – Read the Documentation Section 3 covers agentic applications. Understand how to build multi-step reasoning systems with tool calling.
MLflow Prompt Registry – Read the Documentation Prompt design is tested throughout the exam, and the March 2026 version explicitly tests prompt versioning and lifecycle (Section 4). The Prompt Registry is where you author, version, alias, and promote prompts. For prompt-craft fundamentals, try the AI Playground.
MLflow for GenAI
MLflow 3.0: Build, Evaluate, and Deploy GenAI with Confidence – Read the Blog Post MLflow 3 covers GenAI lifecycle on Databricks: Tracing, Scoring, Prompt Registry, and SME feedback all live here. The 2025 launch post covers the features the March 2026 exam now tests.
Logging Chains and Models – Read the Documentation Understand how to log LangChain chains and custom PyFunc models with MLflow.
Model Serving & Deployment
Foundation Model APIs – Read the Documentation Learn about Databricks’ pay-per-token serving for foundation models – important for Section 4.
Model Serving Endpoints – Read the Documentation Understand endpoint creation, scaling, and access control for serving GenAI applications.
Evaluation & Monitoring
Agent Evaluation – Read the Documentation Section 6 focuses on evaluation. Learn about LLM judges, metrics, and evaluation strategies.
Inference Tables – Read the Documentation Understand how to capture and analyze inference requests and responses for monitoring.
Free Resources
Generative AI Fundamentals – Free Course A ~2-hour background course and recommended starting point for this certification. It’s free, and completing it earns an accredited badge you can share on LinkedIn to show progress.
AI Agent Fundamentals – Free Course A free, ~1.5-hour follow-on that covers AI agents on Databricks, the Mosaic AI platform, Agent Bricks, and multi-agent systems. Worth doing given that Sections 3 and 4 (the agent-heavy half of the exam) are now 52% of the score.
Databricks Free Edition – Sign Up Get hands-on practice with a free Databricks workspace – no credit card required.
📊 Exam Breakdown & Study Strategy
Exam Overview
| Attribute | Value |
|---|---|
| Exam Version | March 18, 2026 |
| Total Scored Questions | 45 |
| Time Limit | 90 minutes |
| Registration Fee | $200 |
| Validity | 2 years |
| Source Guide | Official PDF |
Section Summary
| Section | Topic | Weight | Objectives |
|---|---|---|---|
| 1 | Design Applications | 14% | 6 |
| 2 | Data Preparation | 14% | 8 |
| 3 | Application Development | 30% | 13 |
| 4 | Assembling and Deploying Applications | 22% | 15 |
| 5 | Governance | 8% | 4 |
| 6 | Evaluation and Monitoring | 12% | 10 |
🎯 How to Use This Guide Effectively
I’ve organized resources into four categories for each exam objective. Here’s how I recommend using them:
📚 Official Documentation (docs.databricks.com)
This is where you get the “official” definition and syntax. I use docs as my reference material when I need precise technical details.
My approach:
- Start with the “Getting Started” and “How-to” sections
- Bookmark key pages for quick review before the exam
- Don’t try to read every doc page – use them as reference material when you need specifics
Best for: Understanding exact syntax, parameters, and technical specifications
🎯 Interactive Demos (databricks.com/resources/demos)
Demos are where things click for me. Watching someone navigate the UI helps me understand workflows much faster than reading about them.
How I use demos:
- Before watching: I read the exam objective so I know what to focus on
- During the demo: I take screenshots of important configuration screens
- After the demo: I try to recreate what I saw in my own workspace – this is key!
Demo types:
- Hands-On Tutorials: Step-by-step guides (follow along in your workspace)
- Product Tours: Quick 3-5 minute overviews (watch these first)
- Video Demos: In-depth demonstrations (take notes, then practice)
Best for: Understanding UI workflows and seeing features in action
🎓 Training Resources (Databricks Academy)
If you prefer structured learning paths, these are great resources.
Training Courses (databricks.com/training):
- The official “Generative AI Engineering” path is excellent
- Many self-paced courses are free via Databricks Academy
- Hands-on labs are included – make sure you actually do them!
Best for: Structured learning and understanding how products fit together
My Recommended Study Path
I weight this plan by section weight. Application Development (30%) and Assembling & Deploying (22%) are the big ones, so most of the time goes there.
Week 1: Foundations & Design (Sections 1 + 2, 28%)
- Prompt engineering basics, the AI Playground, and structured-output prompting
- Chunking strategies and the RAG data pipeline
- Vector Search, embeddings, and retrieval evaluation
- When to reach for Agent Bricks (Knowledge Assistant, Multi-Agent Supervisor, Information Extraction) vs the Agent Framework
Week 2-3: Application Development (Section 3, 30%)
- Build a single-agent app with MLflow + the Agent Framework, end-to-end
- Add Agent Bricks options for low-code paths
- Multi-agent patterns with Genie Spaces and conversational API
- LLM and embedding model selection from Foundation Model APIs
- Guardrails via AI Gateway
Week 4: Assembling & Deploying (Section 4, 22%)
- PyFunc / MLflow chain logging and Unity Catalog registration
- Vector Search index configuration (standard vs storage-optimized, hybrid + reranking)
- Serving via Model Serving and Foundation Model APIs
- MCP servers (managed, external, custom) and persistent agent memory
- Prompt Registry lifecycle and CI/CD for agents
- Building a user-facing UI on Databricks Apps
Week 5: Evaluation & Monitoring (Section 6, 12%)
- MLflow Tracing and Scoring on a deployed agent
- Custom Scorers via
mlflow.genai.evaluate() - AI Gateway inference tables, usage tables, and rate limiting
- Incorporating SME feedback with rubrics
Week 6: Governance + practice exams (Section 5, 8%)
- Unity Catalog row filters and column masks for PII
- Legal/licensing checks on data sources
- Practice exam runs and gap-fill
Practice & Validation
Hands-On Practice (this is critical):
- Sign up for Databricks Free Edition (free, no credit card)
- Build at least one agent end-to-end. Sections 3 and 4 together are 52% of the exam, so don’t stop at “I made a RAG”
- Start with Agent Bricks for the no-code path, or the Agent Framework with MLflow if you want to author in Python
- Wire RAG and Vector Search into your agent, not as standalone exercises
- Deploy via Model Serving and ship a user-facing UI on Databricks Apps
- Wrap it in AI Gateway so you can see inference tables, usage tables, and rate limits in practice
- Iterate prompts in the Prompt Registry and evaluate with MLflow Scorers
- Hands-on practice is the difference between passing and actually understanding the platform
Section 1: Design Applications
Section Overview: 6 objectives
Recommended Demos for This Section
Start with these demos to get hands-on experience:
🎓 Hands-On Tutorials (Follow along in your workspace):
🎥 Product Tours (Quick 3-5 minute overviews):
📹 Video Demos (In-depth demonstrations):
1.1 Designing Prompts for Structured Outputs
Objective: Design a prompt that elicits a specifically formatted response
📚 Official Documentation:
Top Demos:
🎓 Training Resources:
- Generative AI Engineering with Databricks (umbrella)
- Building Single-Agent Applications on Databricks
1.2 Selecting Model Tasks for Business Requirements
Objective: Select model tasks to accomplish a given business requirement
📚 Official Documentation:
1.3 Selecting Chain Components
Objective: Select chain components for a desired model input and output
📚 Official Documentation:
1.4 Translating Business Goals to AI Pipelines
Objective: Translate business use case goals into a description of the desired inputs and outputs for the AI pipeline
No curated resources for this objective. See the section overview above for general guidance.
1.5 Defining Tools for Multi-Stage Reasoning
Objective: Define and order tools that gather knowledge or take actions for multi-stage reasoning
📚 Official Documentation:
Top Demos:
1.6 Choosing When to Use Agent Bricks
Objective: Determine how and when to use Agent Bricks (Knowledge Assistant, Multiagent Supervisor, Information Extraction) to solve problems
📚 Official Documentation:
Top Demos:
Section 2: Data Preparation
Section Overview: 8 objectives
Recommended Demos for This Section
Start with these demos to get hands-on experience:
🎓 Hands-On Tutorials (Follow along in your workspace):
📹 Video Demos (In-depth demonstrations):
2.1 Applying Chunking Strategies
Objective: Apply a chunking strategy for a given document structure and model constraints
📚 Official Documentation:
🎓 Training Resources:
2.2 Filtering Content for RAG Quality
Objective: Filter extraneous content in source documents that degrades quality of a RAG application
📚 Official Documentation:
Top Demos:
2.3 Extracting Document Content with Python
Objective: Choose the appropriate Python package to extract document content from provided source data and format
No curated resources for this objective. See the section overview above for general guidance.
2.4 Writing Chunks to Delta Lake
Objective: Define operations and sequence to write given chunked text into Delta Lake tables in Unity Catalog
📚 Official Documentation:
Top Demos:
2.5 Identifying Source Documents for RAG
Objective: Identify needed source documents that provide necessary knowledge and quality for a given RAG application
📚 Official Documentation:
2.6 Evaluating Retrieval Performance
Objective: Use tools and metrics to evaluate retrieval performance
📚 Official Documentation:
2.7 Advanced Chunking Strategies
Objective: Design retrieval systems using advanced chunking strategies
📚 Official Documentation:
2.8 Understanding Re-ranking in Retrieval
Objective: Explain the role of re-ranking in the information retrieval process
📚 Official Documentation:
Section 3: Application Development
Section Overview: 13 objectives
Recommended Demos for This Section
Start with these demos to get hands-on experience:
🎓 Hands-On Tutorials (Follow along in your workspace):
🎥 Product Tours (Quick 3-5 minute overviews):
- Mosaic AI Agent Tool Catalog
- Mosaic AI Agent Framework Evaluation
- DBRX Instruct in Databricks Playground
📹 Video Demos (In-depth demonstrations):
3.1 Selecting GenAI Frameworks (LangChain, etc.)
Objective: Select Langchain/similar tools for use in a Generative AI application
📚 Official Documentation:
Top Demos:
🎓 Training Resources:
3.2 Assessing Response Quality and Safety
Objective: Qualitatively assess responses to identify common issues such as quality and safety
📚 Official Documentation:
3.3 Selecting Chunking Strategy Based on Evaluation
Objective: Select chunking strategy based on model & retrieval evaluation
📚 Official Documentation:
3.4 Augmenting Prompts with User Context
Objective: Augment a prompt with additional context from a user’s input based on key fields, terms, and intents
📚 Official Documentation:
3.5 Adjusting LLM Responses with Prompts
Objective: Create a prompt that adjusts an LLM’s response from a baseline to a desired output
📚 Official Documentation:
3.6 Implementing LLM Guardrails
Objective: Implement LLM guardrails to prevent negative outcomes
📚 Official Documentation:
Top Demos:
3.7 Selecting the Best LLM for an Application
Objective: Select the best LLM based on the attributes of the application to be developed
📚 Official Documentation:
Top Demos:
3.8 Selecting Embedding Model Context Length
Objective: Select an embedding model context length based on source documents, expected queries, and optimization strategy
📚 Official Documentation:
3.9 Selecting Models from Hubs/Marketplaces
Objective: Select a model from a model hub or marketplace for a task based on model metadata/model cards
📚 Official Documentation:
Top Demos:
3.10 Selecting Models Based on Experiment Metrics
Objective: Select the best model for a given task based on common metrics generated in experiments
📚 Official Documentation:
3.11 Using MLflow + Agent Framework for Agentic Systems
Objective: Utilize MLflow and Agent Framework for developing agentic systems
📚 Official Documentation:
Top Demos:
3.12 Comparing Evaluation and Monitoring Phases
Objective: Compare the evaluation and monitoring phases of the Gen AI application life cycle
📚 Official Documentation:
3.13 Multi-Agent Systems with Genie Spaces
Objective: Enable multi-agent systems to leverage Genie Spaces or conversational API to retrieve data
📚 Official Documentation:
Top Demos:
Section 4: Assembling and Deploying Applications
Section Overview: 15 objectives
Recommended Demos for This Section
Start with these demos to get hands-on experience:
🎓 Hands-On Tutorials (Follow along in your workspace):
🎥 Product Tours (Quick 3-5 minute overviews):
📹 Video Demos (In-depth demonstrations):
4.1 Coding Chains with PyFunc Models
Objective: Code a chain using a pyfunc model with pre- and post-processing
📚 Official Documentation:
🎓 Training Resources:
- Generative AI Application Deployment and Monitoring
- Building Single-Agent Applications on Databricks
4.2 Controlling Access to Model Serving Endpoints
Objective: Control access to resources from model serving endpoints
📚 Official Documentation:
Top Demos:
4.3 Coding Simple Chains
Objective: Code a simple chain according to requirements
📚 Official Documentation:
4.4 RAG Application Components
Objective: Choose the basic elements needed to create a RAG application: model flavor, embedding model, retriever, dependencies, input examples, model signature
📚 Official Documentation:
4.5 Registering Models to Unity Catalog
Objective: Register the model to Unity Catalog using MLflow
📚 Official Documentation:
Top Demos:
4.6 Creating and Querying Vector Search Indexes
Objective: Create and query a Vector Search index
📚 Official Documentation:
Top Demos:
4.7 Serving LLMs with Foundation Model APIs
Objective: Identify how to serve an LLM application that leverages Foundation Model APIs
📚 Official Documentation:
Top Demos:
4.8 Mosaic AI Vector Search Concepts
Objective: Explain the key concepts and components of Mosaic AI Vector Search
📚 Official Documentation:
Top Demos:
- Lakehouse Monitoring And Vector Search
- Mosaic AI Agent Tool Catalog
- Mosaic AI Agent Framework Evaluation
4.9 Batch Inference with ai_query()
Objective: Identify batch inference workloads and apply ai_query() appropriately
📚 Official Documentation:
4.10 Configuring Vector Search for Latency and Cost
Objective: Configure vector search for a particular solution based on number of embeddings, update frequency, latency, and cost requirements
📚 Official Documentation:
- Mosaic AI Vector Search
- Create a Vector Search Index
- Query a Vector Search Index (hybrid search & reranking)
4.11 Persistent Datastores for Agent Memory
Objective: Configure a persistent datastore to store and retrieve intermediate memory or structured information
📚 Official Documentation:
4.12 CI/CD Best Practices for Agents
Objective: Apply CI/CD best practices such as updating a Vector Search index, promoting prompts across environments, and testing individual components of an agent
📚 Official Documentation:
4.13 Integrating Managed, External, and Custom MCP Servers
Objective: Integrate managed, external, and custom MCP servers based on given application requirements
📚 Official Documentation:
4.14 Prompt Version Control and Lifecycle
Objective: Apply prompt version control and manage prompt lifecycle
📚 Official Documentation:
4.15 Building Agent UIs (Apps, Slack, Teams)
Objective: Develop an appropriate interactive user facing interface for an agent usage scenario (Apps, Slack, Teams, etc.)
📚 Official Documentation:
Top Demos:
Section 5: Governance
Section Overview: 4 objectives
Recommended Demos for This Section
Start with these demos to get hands-on experience:
🎓 Hands-On Tutorials (Follow along in your workspace):
🎥 Product Tours (Quick 3-5 minute overviews):
5.1 Using Masking Techniques as Guardrails
Objective: Use masking techniques as guard rails to meet a performance objective
📚 Official Documentation:
🎓 Training Resources:
5.2 Protecting Against Malicious Inputs
Objective: Select guardrail techniques to protect against malicious user inputs to a Gen AI application
📚 Official Documentation:
5.3 Managing Legal and Licensing Requirements
Objective: Use legal/licensing requirements for data sources to avoid legal risk
📚 Official Documentation:
5.4 Mitigating Problematic Text in GenAI Sources
Objective: Recommend an alternative for problematic text mitigation in a data source feeding a GenAI application
📚 Official Documentation:
Section 6: Evaluation and Monitoring
Section Overview: 10 objectives
Recommended Demos for This Section
Start with these demos to get hands-on experience:
🎥 Product Tours (Quick 3-5 minute overviews):
📹 Video Demos (In-depth demonstrations):
6.1 Selecting LLMs Based on Evaluation Metrics
Objective: Select an LLM choice (size and architecture) based on a set of quantitative evaluation metrics
📚 Official Documentation:
Top Demos:
🎓 Training Resources:
- Generative AI Application Deployment and Monitoring
- Generative AI Engineering with Databricks (umbrella)
6.2 Selecting Key Metrics for LLM Monitoring
Objective: Select key metrics to monitor for a specific LLM deployment scenario
📚 Official Documentation:
Top Demos:
- Lakehouse Monitoring And Vector Search
- Lakehouse Monitoring Databricks
- Monitor Your Data Quality With Lakehouse Monitoring
6.3 Evaluating Agents with MLflow Scoring and Tracing
Objective: Evaluate agent performance using MLflow scoring and tracing
📚 Official Documentation:
6.4 Using Inference Logging for Performance Assessment
Objective: Use inference logging to assess deployed RAG application performance
📚 Official Documentation:
6.5 Controlling LLM Costs
Objective: Use Databricks features to control LLM costs
📚 Official Documentation:
6.6 Using Inference Tables and Agent Monitoring
Objective: Use inference tables and Agent Monitoring to track a live LLM endpoint
📚 Official Documentation:
Top Demos:
- Lakehouse Monitoring And Vector Search
- Lakehouse Monitoring Databricks
- Monitor Your Data Quality With Lakehouse Monitoring
6.7 Understanding Evaluation Judges and Ground Truth
Objective: Identify evaluation judges that require ground truth
📚 Official Documentation:
6.8 Tracking with AI Gateway
Objective: Use AI Gateway (Inference Tables, Usage Tables, and rate limiting) to track an LLM or agent deployed via Agent Framework
📚 Official Documentation:
Top Demos:
6.9 Custom Scorers for Agent and LLM Evaluation
Objective: Use Databricks custom Scorers for evaluating agents and LLMs
📚 Official Documentation:
6.10 Incorporating SME Feedback
Objective: Incorporate SME feedback to improve agent performance
📚 Official Documentation:
Quick Reference Table
| Objective | Description | Demo Count |
|---|---|---|
| 1.1 | Design a prompt that elicits a specifically formatted respon… | 1 |
| 1.2 | Select model tasks to accomplish a given business requiremen… | 0 |
| 1.3 | Select chain components for a desired model input and output | 0 |
| 1.4 | Translate business use case goals into a description of the … | 0 |
| 1.5 | Define and order tools that gather knowledge or take actions… | 3 |
| 1.6 | Determine how and when to use Agent Bricks (Knowledge Assist… | 3 |
| 2.1 | Apply a chunking strategy for a given document structure and… | 0 |
| 2.2 | Filter extraneous content in source documents that degrades … | 1 |
| 2.3 | Choose the appropriate Python package to extract document co… | 0 |
| 2.4 | Define operations and sequence to write given chunked text i… | 3 |
| 2.5 | Identify needed source documents that provide necessary know… | 0 |
| 2.6 | Use tools and metrics to evaluate retrieval performance | 0 |
| 2.7 | Design retrieval systems using advanced chunking strategies | 0 |
| 2.8 | Explain the role of re-ranking in the information retrieval … | 0 |
| 3.1 | Select Langchain/similar tools for use in a Generative AI ap… | 3 |
| 3.2 | Qualitatively assess responses to identify common issues suc… | 0 |
| 3.3 | Select chunking strategy based on model & retrieval evaluati… | 0 |
| 3.4 | Augment a prompt with additional context from a user’s input… | 0 |
| 3.5 | Create a prompt that adjusts an LLM’s response from a baseli… | 0 |
| 3.6 | Implement LLM guardrails to prevent negative outcomes | 3 |
| 3.7 | Select the best LLM based on the attributes of the applicati… | 3 |
| 3.8 | Select an embedding model context length based on source doc… | 0 |
| 3.9 | Select a model from a model hub or marketplace for a task ba… | 2 |
| 3.10 | Select the best model for a given task based on common metri… | 0 |
| 3.11 | Utilize MLflow and Agent Framework for developing agentic sy… | 3 |
| 3.12 | Compare the evaluation and monitoring phases of the Gen AI a… | 0 |
| 3.13 | Enable multi-agent systems to leverage Genie Spaces or conve… | 3 |
| 4.1 | Code a chain using a pyfunc model with pre- and post-process… | 0 |
| 4.2 | Control access to resources from model serving endpoints | 2 |
| 4.3 | Code a simple chain according to requirements | 0 |
| 4.4 | Choose the basic elements needed to create a RAG application… | 0 |
| 4.5 | Register the model to Unity Catalog using MLflow | 3 |
| 4.6 | Create and query a Vector Search index | 1 |
| 4.7 | Identify how to serve an LLM application that leverages Foun… | 3 |
| 4.8 | Explain the key concepts and components of Mosaic AI Vector … | 3 |
| 4.9 | Identify batch inference workloads and apply ai_query() appr… | 0 |
| 4.10 | Configure vector search for a particular solution based on n… | 0 |
| 4.11 | Configure a persistent datastore to store and retrieve inter… | 0 |
| 4.12 | Apply CI/CD best practices such as updating a Vector Search … | 0 |
| 4.13 | Integrate managed, external, and custom MCP servers based on… | 0 |
| 4.14 | Apply prompt version control and manage prompt lifecycle | 0 |
| 4.15 | Develop an appropriate interactive user facing interface for… | 1 |
| 5.1 | Use masking techniques as guard rails to meet a performance … | 0 |
| 5.2 | Select guardrail techniques to protect against malicious use… | 0 |
| 5.3 | Use legal/licensing requirements for data sources to avoid l… | 0 |
| 5.4 | Recommend an alternative for problematic text mitigation in … | 0 |
| 6.1 | Select an LLM choice (size and architecture) based on a set … | 3 |
| 6.2 | Select key metrics to monitor for a specific LLM deployment … | 3 |
| 6.3 | Evaluate agent performance using MLflow scoring and tracing | 0 |
| 6.4 | Use inference logging to assess deployed RAG application per… | 0 |
| 6.5 | Use Databricks features to control LLM costs | 0 |
| 6.6 | Use inference tables and Agent Monitoring to track a live LL… | 3 |
| 6.7 | Identify evaluation judges that require ground truth | 0 |
| 6.8 | Use AI Gateway (Inference Tables, Usage Tables, and rate lim… | 1 |
| 6.9 | Use Databricks custom Scorers for evaluating agents and LLMs | 0 |
| 6.10 | Incorporate SME feedback to improve agent performance | 0 |
Study Resources
Official Training (Databricks Academy)
The March 2026 exam guide names the Generative AI Engineering with Databricks pathway as the primary preparation path. It bundles four self-paced modules:
- Generative AI Engineering with Databricks (umbrella)
- Building Retrieval Agents on Databricks
- Building Single-Agent Applications on Databricks
- Generative AI Application Deployment and Monitoring
The exam guide also names a Generative AI Application Evaluation and Governance module. It has no working standalone catalog URL at present; you reach it via the umbrella pathway above.
For background, Generative AI Fundamentals is free.
Certification Information
- Generative AI Engineer Associate Exam Page
- Databricks Free Edition – Practice for free
Key Documentation
- RAG on Databricks – Building RAG systems
- Mosaic AI Vector Search – Embeddings, indexing, hybrid search & reranking
- Agent Framework – Authoring custom agents
- Agent Bricks – Knowledge Assistant, Multi-Agent Supervisor, Information Extraction
- Foundation Model APIs – LLM and embedding model serving
- MLflow Models – Logging chains, PyFunc models
- MLflow 3 GenAI Evaluation & Monitoring – LLM judges, custom Scorers, SME feedback
- MLflow Tracing – Agent observability
- MLflow Prompt Registry – Prompt versions, aliases, lifecycle
- Unity AI Gateway – Inference Tables, Usage Tables, rate limiting, guardrails
- Model Context Protocol (MCP) – Managed, external, and custom MCP servers
- Databricks Apps – Building user-facing agent UIs
Last Updated: May 12, 2026