Databricks Certified Machine Learning Associate
AI and Data Foundations
Review the AI, machine learning, data, and generative AI concepts that appear across the exam.
Official Scope and Verification
This lesson is mapped to the verified Databricks Certified Machine Learning Associate outline. Official sources and public status were rechecked on 2026-07-13. Provider pages remain authoritative for late-breaking blueprint, availability, scheduling, price, language, delivery, and retake changes.
Current Databricks proctored certification with published domain percentages.
Official Objectives Emphasized Here
| Domain or objective area | Published weight | Key objective groups | Official source |
|---|---|---|---|
| Databricks Machine Learning | 38% | Identify the best practices of an MLOps strategy; Identify the advantages of using ML runtimes; Identify how AutoML facilitates model and feature selection; Identify the advantages AutoML brings to the model development process; Identify benefits of account-level Unity Catalog feature store tables over workspace-level tables; Create a feature store table in Unity Catalog; Write data to a feature store table; Train a model with features from a feature store table; Score a model using features from a feature store table; Describe the differences between online and offline feature tables; Identify the best run using the MLflow Client API; Manually log metrics, artifacts, and models in an MLflow Run; Identify information available in the MLflow UI; Register a model using the MLflow Client API in the Unity Catalog registry; Identify benefits of registering models in the Unity Catalog registry over the workspace registry; Identify when promoting code is preferred over promoting models and vice versa; Set or remove a tag for a model; Promote a challenger model to a champion model using aliases | Databricks official Machine Learning Associate exam guide PDF |
| Data Processing | 19% | Compute summary statistics on a Spark DataFrame using .summary() or dbutils data summaries; Remove outliers from a Spark DataFrame based on standard deviation or IQR; Create visualizations for categorical or continuous features; Compare two categorical or two continuous features using the appropriate method; Compare and contrast imputing missing values with the mean, median, or mode; Impute missing values with the mode, mean, or median value; Use one-hot encoding for categorical features; Identify model types or data sets where one-hot encoding is or is not appropriate; Identify scenarios where log scale transformation is appropriate | Databricks official Machine Learning Associate exam guide PDF |
| Model Development | 31% | Use ML foundations to select the appropriate algorithm for a given model scenario; Identify methods to mitigate data imbalance in training data; Compare estimators and transformers; Develop a training pipeline; Use Hyperopt fmin to tune model hyperparameters; Perform random, grid, or Bayesian search for hyperparameter tuning; Parallelize single node models for hyperparameter tuning; Describe benefits and downsides of cross-validation over a train-validation split; Perform cross-validation as part of model fitting; Identify the number of models trained in grid-search and cross-validation; Use common classification metrics such as F1, log loss, and ROC/AUC; Use common regression metrics such as RMSE, MAE, and R-squared; Choose the most appropriate metric for a scenario objective; Exponentiate log-transformed variables before calculating metrics or interpreting predictions; Assess the impact of model complexity and the bias-variance tradeoff on model performance | Databricks official Machine Learning Associate exam guide PDF |
| Model Deployment | 12% | Identify differences and advantages of batch, realtime, and streaming model serving approaches; Deploy a custom model to a model endpoint; Use pandas to perform batch inference; Identify how streaming inference is performed with Delta Live Tables; Deploy and query a model for realtime inference; Split data between endpoints for realtime inference | Databricks official Machine Learning Associate exam guide PDF |
Authoritative Sources for This Scope
- Databricks official Machine Learning Associate exam guide PDF - Official source; accessed 2026-07-13.
This module gives you the baseline AI and data language needed for Databricks Certified Machine Learning Associate. The goal is not to become a research scientist. The goal is to read an official learning or assessment scenario and know which concept is being tested.
Core Concepts To Know
- AI versus ML versus GenAI. AI is the broad goal of useful machine behavior. ML learns patterns from data. GenAI creates or transforms content such as text, code, images, audio, or structured summaries.
- Training versus inference. Training builds or adapts behavior from data. Inference uses a trained model to produce an output for a new input.
- Prediction versus generation. Prediction chooses a label, score, class, or forecast. Generation creates new content and must be checked for grounding, safety, and quality.
- Foundation model. A large pretrained model that can be adapted through prompting, retrieval, fine-tuning, tools, or workflow design.
- Embedding. A numeric representation of meaning that helps search, clustering, recommendations, semantic similarity, and RAG.
- Evaluation. The discipline of measuring whether outputs are correct, useful, safe, fair, and stable enough for the use case.
Data Foundations
Most AI failures start with data assumptions. For Databricks scenarios, ask where the data comes from, who is allowed to use it, whether it is current, whether labels are reliable, and whether sensitive information is protected.
| Data issue | Why it is tested | Self-learner check |
|---|---|---|
| Missing or stale data | The model may answer confidently from incomplete evidence. | Ask whether retrieval, refresh, or data validation is needed. |
| Biased or unrepresentative data | The output can treat groups or edge cases unfairly. | Look for fairness testing, representative samples, and human review. |
| Sensitive data | Prompts, files, logs, and model outputs can expose private or regulated information. | Apply classification, access control, encryption, masking, and retention limits. |
| Poor labels or definitions | A model cannot learn or evaluate a target that the organization has not defined clearly. | Define success metrics before choosing the model or tool. |
Model And Workflow Vocabulary
- Prompting: giving the model a task, context, constraints, examples, and desired output format.
- Grounding: connecting the model to trusted source material so outputs are tied to current facts.
- RAG: retrieving relevant content and passing it to the model at response time, often better than fine-tuning when source material changes frequently.
- Fine-tuning: adapting a model with training examples, useful for repeatable style or task behavior but not a replacement for current source retrieval.
- Agents: systems that plan or call tools to complete tasks; they need boundaries, permissions, logs, and fallback behavior.
- Human oversight: review by a person when the output affects safety, money, legal rights, employment, healthcare, education, or other high-impact decisions.
Provider-Specific Lens
For Databricks Certified Machine Learning Associate, tie every AI concept back to lakehouse AI, ML engineering, generative AI, context engineering, and data governance. A generic definition is useful only if you can apply it to a scenario from Databricks.
- Unity Catalog
- MLflow
- Model Serving
- Vector Search
- Mosaic AI
- Lakehouse monitoring
Track-Specific Vocabulary Priorities
- Read the exact credential title first. Many AI credentials are role-based, so the same AI concept can be tested differently for an engineer, architect, auditor, business leader, teacher, or administrator.
- Translate every objective into a real scenario with a user, data source, risk constraint, and expected output.
- Separate durable AI principles from provider product names so you can still reason when a product name changes.
- Connect supervised learning, unsupervised learning, feature handling, model selection, validation, deployment, and drift monitoring.
- Treat data quality, leakage, label definition, and evaluation design as first-class exam topics.
- Know when an experiment, notebook, pipeline, model registry, endpoint, or monitoring control is the next logical step.
Example: RAG Or Fine-Tuning
Scenario: a support team needs answers from policy documents that change every month. The best first pattern is usually retrieval-grounded generation because the answer should come from current documents. Fine-tuning may help style or task behavior, but it does not automatically keep the model synchronized with the latest policy.
Common trap: choosing the more advanced-sounding option instead of the pattern that matches the data-change requirement.
Practice Routine
- Make flashcards for the vocabulary above, but put the definition on one side and a workplace example on the other.
- For every provider tool you study, write the AI concept it maps to: search, classification, generation, orchestration, monitoring, governance, or security.
- When you miss a question, classify the miss as vocabulary, data, model choice, security, or operations. Review the category, not just that one answer.
Useful Links
- Databricks Certification and Badging - Official Databricks certification and accreditation catalog.
- Databricks Academy - Official learning platform entry point.
- NIST AI Risk Management Framework - General reference for trustworthy AI risk management.