CertSafari

    Databricks Certified Generative AI Engineer Associate Lessons

    56 lessons, one per exam-guide subdomain, in the order the guide teaches them. Every claim is cited to the official documentation.

    Domain 1: Design Applications

    6 lessons · 11% of the exam

    1. 1.1Prompting for formatted LLM output: instructions, few-shot examples and structured outputs

      Subdomain 1.1: Design a prompt that elicits a specifically formatted response

      17 min read

    2. 1.2Choosing a Model Task for a Business Requirement on Databricks

      Subdomain 1.2: Select model tasks to accomplish a given business requirement

      2 pages · 18 min read

    3. 1.3RAG Chain Components: Shaping What Reaches the LLM

      Subdomain 1.3: Select chain components for a desired model input and output

      2 pages · 20 min read

    4. 1.4Defining GenAI Pipeline Inputs and Outputs from Business Goals

      Subdomain 1.4: Translate business use case goals into a description of the desired inputs and outputs for the AI pipeline

      2 pages · 21 min read

    5. 1.5Agent Design Patterns: Who Decides Tool Order in Multi-Stage Reasoning

      Subdomain 1.5: Define and order tools that gather knowledge or take actions for multi-stage reasoning

      2 pages · 18 min read

    6. 1.6Agent Bricks: Knowledge Assistant vs Information Extraction

      Subdomain 1.6: Determine how and when to use Agent Bricks (Knowledge Assistant, Multiagent Supervisor, Information Extraction) to solve problems

      2 pages · 16 min read

    Domain 2: Data Preparation

    8 lessons · 14% of the exam

    1. 2.1Chunking Strategies for Document Structure and Model Limits

      Subdomain 2.1: Apply a chunking strategy for a given document structure and model constraints

      2 pages · 18 min read

    2. 2.2Filtering Extraneous Content from RAG Source Documents

      Subdomain 2.2: Filter extraneous content in source documents that degrades quality of a RAG application

      15 min read

    3. 2.3Choosing a Python Package to Parse PDF, HTML, DOCX and Images

      Subdomain 2.3: Choose the appropriate Python package to extract document content from provided source data and format.

      16 min read

    4. 2.4Designing a Unity Catalog Delta Table for Chunked Text

      Subdomain 2.4: Define operations and sequence to write given chunked text into Delta Lake tables in Unity Catalog

      2 pages · 20 min read

    5. 2.5Choosing Source Documents for a RAG Application

      Subdomain 2.5: Identify needed source documents that provide necessary knowledge and quality for a given RAG application

      11 min read

    6. 2.6Retrieval Metrics: Precision, Recall, DCG@10 and NDCG

      Subdomain 2.6: Use tools and metrics to evaluate retrieval performance

      2 pages · 18 min read

    7. 2.7RAG Chunking Strategies: Fixed-Size, Paragraph, Format-Specific and Semantic

      Subdomain 2.7: Design retrieval systems using advanced chunking strategies

      2 pages · 22 min read

    8. 2.8Re-ranking in RAG Retrieval: Role, Cross-Encoders and Trade-offs

      Subdomain 2.8: Explain the role of re-ranking in the information retrieval process

      2 pages · 17 min read

    Domain 3: Application Development

    13 lessons · 23% of the exam

    1. 3.1Choosing a Design Pattern and Agent Framework (LangChain, DSPy, and Similar)

      Subdomain 3.1: Select Langchain/similar tools for use in a Generative AI application.

      2 pages · 20 min read

    2. 3.2Quality and Safety Issues in GenAI Responses: What to Look For

      Subdomain 3.2: Qualitatively assess responses to identify common issues such as quality and safety

      2 pages · 21 min read

    3. 3.3RAG Chunking Strategies: Size, Overlap, Structure and Model Limits

      Subdomain 3.3: Select chunking strategy based on model & retrieval evaluation

      2 pages · 17 min read

    4. 3.4Augmenting Prompts with User Intents, Key Fields and Terms

      Subdomain 3.4: Augment a prompt with additional context from a user's input based on key fields, terms, and intents

      15 min read

    5. 3.5Prompt Engineering: Moving an LLM from Baseline to Desired Output

      Subdomain 3.5: Create a prompt that adjusts an LLM's response from a baseline to a desired output

      2 pages · 20 min read

    6. 3.6LLM Guardrails on Databricks: Service Policies, Phases and Built-in Checks

      Subdomain 3.6: Implement LLM guardrails to prevent negative outcomes

      2 pages · 20 min read

    7. 3.7Choosing an LLM by Capability Fit: Task, Modality, Tier and Constraints

      Subdomain 3.7: Select the best LLM based on the attributes of the application to be developed

      2 pages · 16 min read

    8. 3.8Embedding Model Context Length: Fitting Documents and Queries

      Subdomain 3.8: Select an embedding model context length based on source documents, expected queries, and optimization strategy

      2 pages · 16 min read

    9. 3.9Finding Foundation Models in system.ai and Databricks Marketplace

      Subdomain 3.9: Select a model from a model hub or marketplace for a task based on model metadata/model cards

      2 pages · 18 min read

    10. 3.10Comparing GenAI App Versions with MLflow Evaluation Runs

      Subdomain 3.10: Select the best model for a given task based on common metrics generated in experiments

      2 pages · 22 min read

    11. 3.11Authoring Agents with MLflow ResponsesAgent and Agent Framework

      Subdomain 3.11: Utilize MLflow and Agent Framework for developing agentic systems

      2 pages · 24 min read

    12. 3.12GenAI Evaluation Phase: Scoring an Agent Before Release

      Subdomain 3.12: Compare the evaluation and monitoring phases of the Gen AI application life cycle

      2 pages · 19 min read

    13. 3.13Genie Agents as a Structured-Data Tool: Genie API and the Genie MCP Server

      Subdomain 3.13: Enable multi-agent systems to leverage Genie Spaces or conversational API to retrieve data

      2 pages · 19 min read

    Domain 4: Assembling and Deploying Applications

    15 lessons · 27% of the exam

    1. 4.1Pyfunc Chains: load_context, predict, and Pre/Post-Processing Hooks

      Subdomain 4.1: Code a chain using a pyfunc model with pre- and post-processing

      2 pages · 18 min read

    2. 4.2Model Serving Endpoint Identity, Grants and Query Permissions

      Subdomain 4.2: Control access to resources from model serving endpoints

      2 pages · 22 min read

    3. 4.3RAG Chain Steps: Building a Simple Deterministic Chain

      Subdomain 4.3: Code a simple chain according to requirements

      2 pages · 19 min read

    4. 4.4RAG Application Building Blocks: Flavor, Embeddings, Retriever, Signature

      Subdomain 4.4: Choose the basic elements needed to create a RAG application: model flavor, embedding model, retriever, dependencies, input examples, model signature

      16 min read

    5. 4.5Register a Model to Unity Catalog with MLflow

      Subdomain 4.5: Register the model to Unity Catalog using MLflow

      2 pages · 21 min read

    6. 4.6Create a Vector Search (AI Search) Endpoint and Index on Databricks

      Subdomain 4.6: Create and query a Vector Search index

      2 pages · 20 min read

    7. 4.7Foundation Model APIs: Serving Modes and Provisioned Throughput Endpoints

      Subdomain 4.7: Identify how to serve an LLM application that leverages Foundation Model APIs

      2 pages · 21 min read

    8. 4.8Mosaic AI Vector Search: Endpoints, Indexes, Embeddings and Query Types

      Subdomain 4.8: Explain the key concepts and components of Mosaic AI Vector Search

      17 min read

    9. 4.9Batch inference workloads: when to use ai_query on Databricks

      Subdomain 4.9: Identify batch inference workloads and apply ai_query() appropriately

      2 pages · 20 min read

    10. 4.10Vector Search Endpoint Sizing: Standard vs Storage-Optimized by Scale, Latency and Cost

      Subdomain 4.10: Configure vector search for a particular solution based on number of embeddings, update frequency, latency, and cost requirements.

      2 pages · 17 min read

    11. 4.11Agent Sessions on Lakebase: Persisting Conversation State

      Subdomain 4.11: Configure a persistent datastore to store and retrieve intermediate memory or structured information.

      2 pages · 23 min read

    12. 4.12Promoting Prompts Across Environments with MLflow Prompt Registry Aliases

      Subdomain 4.12: Apply CI/CD best practices such as updating a Vector Search index, promoting prompts across environments, and testing individual components of an agent.

      2 pages · 19 min read

    13. 4.13Integrating Managed, External, and Custom MCP Servers on Databricks

      Subdomain 4.13: Integrate managed, external, and custom MCP servers based on a given application requirements

      18 min read

    14. 4.14MLflow Prompt Registry: Prompt Versioning, Aliases and Lifecycle

      Subdomain 4.14: Apply prompt version control and manage prompt lifecycle

      16 min read

    15. 4.15Databricks Apps Chat UI for Agents: Template, Querying, and Authorization

      Subdomain 4.15: Develop an appropriate interactive user facing interface for an agent usage scenario (Apps, Slack, Teams, etc.)

      2 pages · 24 min read

    Domain 5: Governance

    4 lessons · 7% of the exam

    1. 5.1Column Masks as Guardrails: Masking PII Without Losing Performance

      Subdomain 5.1: Use masking techniques as guard rails to meet a performance objective

      13 min read

    2. 5.2Built-in Guardrails Against Malicious Prompts on Databricks

      Subdomain 5.2: Select guardrail techniques to protect against malicious user inputs to a Gen AI application

      2 pages · 20 min read

    3. 5.3Data Source Licensing for GenAI Applications

      Subdomain 5.3: Use legal/licensing requirements for data sources to avoid legal risk

      10 min read

    4. 5.4Mitigating Problematic Text in GenAI Data Sources

      Subdomain 5.4: Recommend an alternative for problematic text mitigation in a data source feeding a GenAI application

      15 min read

    Domain 6: Evaluation and Monitoring

    10 lessons · 18% of the exam

    1. 6.1LLM evaluation metrics for model selection: quality, cost and latency

      Subdomain 6.1: Select an LLM choice (size and architecture) based on a set of quantitative evaluation metrics

      2 pages · 19 min read

    2. 6.2LLM Monitoring Metrics: Quality, Retrieval, Cost and Latency

      Subdomain 6.2: Select key metrics to monitor for a specific LLM deployment scenario

      2 pages · 18 min read

    3. 6.3MLflow Tracing and the mlflow.genai.evaluate() Harness

      Subdomain 6.3: Evaluate agent performance using MLflow scoring and tracing

      2 pages · 23 min read

    4. 6.4Inference Tables for Deployed RAG Apps: What Gets Logged

      Subdomain 6.4: Use inference logging to assess deployed RAG application performance

      2 pages · 15 min read

    5. 6.5Unity Gateway Budgets and Rate Limits for LLM Cost Control

      Subdomain 6.5: Use Databricks features to control LLM costs

      2 pages · 18 min read

    6. 6.6Inference Tables: Logging a Live LLM or Agent Endpoint

      Subdomain 6.6: Use inference tables and Agent Monitoring to track a live LLM endpoint

      2 pages · 21 min read

    7. 6.7Built-in LLM Judges That Require Ground Truth

      Subdomain 6.7: Identify evaluation judges that require ground truth

      2 pages · 17 min read

    8. 6.8AI Gateway for Agent Endpoints: Inference Tables, Usage Tables, Rate Limits

      Subdomain 6.8: Use AI Gateway (Inference Tables, Usage Tables, and rate limiting) to track an LLM or agent deployed via Agent Framework.

      15 min read

    9. 6.9Custom Code-Based Scorers in MLflow: Signature, Inputs and Return Values

      Subdomain 6.9: Use Databricks custom Scorers for evaluating agents and LLMs

      2 pages · 21 min read

    10. 6.10Collect SME Feedback with the MLflow Review App

      Subdomain 6.10: Incorporate SME feedback to improve agent performance

      2 pages · 20 min read

    Practice Databricks Certified Generative AI Engineer Associate in quiz mode