About Course
A complete course for the Databricks Generative AI Engineer Associate exam, 100% free, with no paywalls, subscriptions or ads. It follows the Ambush Method: short daily study sessions, the traps each exam topic sets, and exam-style questions with an explanation for every option.
The exam doesn’t ask you to define terms. It hands you a scenario and four plausible answers, and asks what you would do. This course trains that skill chapter by chapter, from choosing a model and preparing documents to deploying, evaluating and monitoring a generative AI application on Databricks.
What’s inside
- An introduction to the exam and the study method, followed by 16 chapters built on the official exam objectives
- Trap alerts, exam tips and a micro-checkpoint in every block
- Recall cards (flashcards) that save your progress to your account
- The 63 exam traps in one line each, and the 15 named mental models that settle a question before you read its options
- A full 45-question, 90-minute mock exam with an explanation for every option and the chapter to revisit
Study 30 minutes to an hour a day.
Course Content
Chapter 0: The Exam, Deconstructed
-
Before We Start
-
What This Chapter Covers
-
The Exam You’re Actually Taking
-
How This Book Works
-
The Method: Study Slow
-
What This Book Asks of You
-
What’s Coming Next
Chapter 1: Choosing the Machine
-
Before We Start
-
What This Chapter Covers
-
What an LLM Actually Is
-
The Task Taxonomy
-
Reading a Model Card Like a Contract
-
The Selection Stack, Assembled
-
What to Take to the Exam
-
What’s Coming Next
Chapter 2: The Prompt Contract
-
Before We Start
-
What This Chapter Covers
-
The Anatomy of a Prompt
-
Forcing the Format
-
Steering from a Baseline
-
Augmenting with the User’s Input
-
What to Take to the Exam
-
What’s Coming Next
Chapter 3: Home Field
-
Before We Start
-
What This Chapter Covers
-
The Managed Workspace
-
The Address of Everything
-
The Lifecycle Spine
-
Foundation Model APIs: Your First API Call
-
The Floor Plan
-
What to Take to the Exam
-
What’s Coming Next
Chapter 4: Data Preparation Without Regret
-
Before We Start
-
What This Chapter Covers
-
Choosing the Sources
-
The Extraction Toolbox
-
Filtering the Garbage
-
Licenses, PII, and Poisoned Text
-
Landing in Delta, Addressed by Unity Catalog
-
What to Take to the Exam
-
What’s Coming Next
Chapter 5: Cutting Documents Without Killing Meaning
-
Before We Start
-
What This Chapter Covers
-
Why Chunk at All, and the Arithmetic
-
The Strategy Menu as a Decision
-
Advanced Strategies
-
The Chunking Pipeline in Code
-
What to Take to the Exam
-
What’s Coming Next
Chapter 6: Embeddings and Mosaic AI Vector Search
-
Before We Start
-
What This Chapter Covers
-
Embeddings, Anchored
-
Choosing the Embedding Model
-
Mosaic AI Vector Search, Anatomically
-
Create, Sync, Query
-
Sizing the Thing
-
What to Take to the Exam
-
What’s Coming Next
Chapter 7: Chains from Business Goal to Code
-
Before We Start
-
What This Chapter Covers
-
From Business Goal to I/O Spec
-
Selecting the Stations
-
LangChain and Its Rivals
-
Coding the Chain, Then Wrapping It
-
What to Take to the Exam
-
What’s Coming Next
Chapter 8: Assembling and Tuning the RAG Application
-
Before We Start
-
What This Chapter Covers
-
The RAG Bill of Materials
-
Measuring Retrieval
-
Re-ranking
-
Closing the Chunking Loop
-
What to Take to the Exam
-
What’s Coming Next
Chapter 9: Agents, Tools, and Memory
-
Before We Start
-
What This Chapter Covers
-
From Chain to Agent
-
Designing and Ordering the Tools
-
Agent Framework and MLflow
-
Memory That Survives the Turn
-
What to Take to the Exam
-
What’s Coming Next
Chapter 10: MCP Servers, Genie Spaces, and Agent Bricks
-
Before We Start
-
What This Chapter Covers
-
MCP: The Tool Protocol
-
Genie as the Structured-Data Specialist
-
Agent Bricks
-
Multi-Agent Patterns
-
What to Take to the Exam
-
What’s Coming Next
Chapter 11: Safety Before Shipping
-
Before We Start
-
What This Chapter Covers
-
Reading Outputs Like an Auditor
-
The Attack Surface
-
Masking as a Guardrail
-
Implementing Guardrails on Databricks
-
What to Take to the Exam
-
What’s Coming Next
Chapter 12: Evaluation That Survives Contact
-
Before We Start
-
What This Chapter Covers
-
Why Evaluation Is Its Own Discipline
-
Choosing a Model by the Numbers
-
MLflow Evaluation: Scoring and Tracing
-
Judges, and Who Needs Ground Truth
-
Custom Scorers and the Human Layer
-
What to Take to the Exam
-
What’s Coming Next
Chapter 13: Serving, Access Control, and Interfaces
-
Before We Start
-
What This Chapter Covers
-
From Registry to Endpoint
-
Access Control at the Endpoint
-
Real-Time and Batch
-
The Human Surface
-
What to Take to the Exam
-
What’s Coming Next
Chapter 14: CI/CD and the Prompt Lifecycle
-
Before We Start
-
What This Chapter Covers
-
Prompts as Release Artifacts
-
What CI Can Promise, and What It Cannot
-
One Release, Walked and Unwound
-
What to Take to the Exam
-
What’s Coming Next
Chapter 15: Monitoring, AI Gateway, and Cost
-
Before We Start
-
What This Chapter Covers
-
What Changes When the Inputs Stop Being Yours
-
The Record You Never Asked For
-
The Place Every Call Already Passes Through
-
What to Watch, and What Watching Costs
-
What to Take to the Exam
-
What’s Coming Next
Chapter 16: Exam Strategy and the Full Mock
-
Before We Start
-
What This Chapter Covers
-
How a Question Is Built
-
Sixty-Three Traps, One Line Each
-
Fifteen Models, and What Each One Decides
-
The Day Itself
-
What to Take to the Exam
-
Full Mock Exam (45 Questions)