Free Databricks Exam Questions

Core Databricks Machine Learning Technologies:

• Databricks Machine Learning - ML runtimes, AutoML, MLOps best practices, and machine learning capabilities
• AutoML - Model and feature selection, advantages in model development process
• Unity Catalog - Feature store tables at account level vs workspace level, creating and managing feature tables
• Feature Store - Online and offline feature tables, writing data, training and scoring models with features
• MLflow Client API - Logging metrics, artifacts, and models, registering models, model registry management
• Unity Catalog Registry - Benefits over workspace registry, model promotion, aliases, tags, champion/challenger models
• Spark DataFrames - Summary statistics, outlier removal, data visualization, feature comparison
• Feature Engineering - Imputation (mean, median, mode), one-hot encoding, log scale transformation
• Model Development - Algorithm selection, data imbalance mitigation, estimators vs transformers, training pipelines
• Hyperparameter Tuning - Hyperopt fmin operation, random search, grid search, Bayesian search, parallelization
• Cross-Validation - Train-validation split comparison, cross-validation implementation, grid-search with cross-validation
• Model Evaluation Metrics - Classification (F1, Log Loss, ROC/AUC), Regression (RMSE, MAE, R-squared), metric selection
• Model Deployment - Batch inference, realtime inference, streaming inference with Delta Live Tables
• Model Serving Endpoints - Custom model deployment, endpoint querying, data splitting between endpoints

Exam Sections (4 Main Domains):

1. Databricks Machine Learning - MLOps best practices, ML runtimes, AutoML, feature stores in Unity Catalog, MLflow Client API, model registry, champion/challenger models
2. Data Processing - Summary statistics, outlier removal, data visualization, feature comparison, imputation, one-hot encoding, log scale transformation
3. Model Development - Algorithm selection, data imbalance mitigation, estimators vs transformers, training pipelines, hyperparameter tuning with Hyperopt, cross-validation, model evaluation metrics, model complexity and bias-variance tradeoff
4. Model Deployment - Batch inference, realtime inference, streaming inference with Delta Live Tables, model serving endpoints, custom model deployment

Foundation Skills Tested:

• Using Databricks to perform basic machine learning tasks
• Understanding and using Databricks machine learning capabilities like AutoML, Unity Catalog, and MLflow
• Exploring data and performing feature engineering
• Building models through training, tuning, and evaluation and selection
• Deploying machine learning models (batch, realtime, and streaming)
• Working with feature stores and Unity Catalog
• Managing model lifecycle with MLflow and Unity Catalog registry
• Applying MLOps best practices