Chapter 3 of 6 · Databricks Generative AI Engineer Associate
Application Development
Model choice, chunking, prompts, guardrails, agents with MLflow and Genie tools.
- 4topics
- 20 minfast track
- 30%of the exam
Topics in this chapter
Take them in order, or jump to the one you need.
- 1Selecting LLMs and ModelsDatabricks GenAI Engineer: choose an LLM by context window, latency, cost and license, read model cards in a hub or Marketplace…5 min
- 2Chunking Strategy and Embedding Context LengthDatabricks GenAI Engineer: pick a chunking strategy by document type and retrieval results, and an embedding model whose context…5 min
- 3Prompts, Context Augmentation and GuardrailsDatabricks GenAI Engineer: augment prompts from user intent and key fields, steer output to a target, spot quality and safety…5 min
- 4Agent Frameworks, MLflow and Genie in Multi Agent SystemsDatabricks GenAI Engineer: choose LangChain style tools, build agents with MLflow and Agent Framework, evaluation vs monitoring…5 min
About this chapter
What this chapter covers
Application Development is 30% of the Databricks Generative AI Engineer Associate exam, the largest chapter by far. It covers how you build the working parts of a generative AI application once the design is settled.
The chapter has four parts: choosing the LLM and embedding model, choosing a chunking strategy and embedding context length, writing and augmenting prompts with guardrails around them, and building agents with MLflow, Mosaic AI Agent Framework and Genie spaces as tools.
Questions are mostly short scenarios in which the stated requirement (latency, cost, freshness, who may access what, quality signal) decides the answer. They are single-answer and some multiple-answer multiple choice, and short Python or SQL is shown as text. There are no exhibits.
How the topics fit together
You first pick an LLM and an embedding model that fit the application. The embedding model's context length then limits how large chunks can be, prompts turn retrieved chunks into a good answer, and the agent ties models, tools and data together under MLflow.
- Model questions ask which LLM, hub model or experiment run fits a stated requirement.
- Chunking questions ask which split suits a document type and what retrieval evaluation shows.
- Prompt questions ask how to move a baseline answer to the desired format, tone or content.
- Guardrail questions ask how to block unsafe or off-topic input and output.
- Agent questions ask which framework, MLflow call or Genie integration fits the task.
Read the requirement word first. "Lowest latency", "no ground truth", "structured data in tables" and "must not leave the workspace" each point to a different component, and the distractors are usually right for some other requirement.
How to use these pages
Each topic page has the same shape, so you can skim for what you need.
- What it is: the idea in plain words.
- How it works: the mechanism, with tables and examples.
- What to remember for the exam: the facts to memorize and how questions are phrased.
- Common mistakes: the classic traps.
- Check yourself: three questions to test your understanding.
Ready for more practice?
The full Databricks Generative AI Engineer Associate question bank has 600 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by Databricks. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.