Machine Learning Associate
Databricks Machine Learning Associate practice questions and study guide
Learn to prepare data, build, tune, evaluate, track and deploy machine learning models on Databricks through short, hands-on lessons that follow the four objective areas of the Machine Learning Associate exam, including AutoML, Unity Catalog feature tables, MLflow and model serving. Independent study material. Not affiliated with or endorsed by Databricks, Inc.; the exam name is a trademark of Databricks, Inc.
What you will learn
- Use ML runtimes, AutoML, Unity Catalog feature tables and MLflow tracking and registry aliases
- Clean and transform data in Spark: summary statistics, outliers, imputation, one-hot encoding and log scale
- Train, tune and validate models with pipelines, grid and Bayesian search, Hyperopt and cross-validation
- Pick the right metric and the right serving approach: batch, streaming or real-time endpoints
Topics by exam domain
MLOps, ML runtimes and AutoML
The platform foundations: MLOps practices, ML runtimes and what AutoML automates.
Feature tables in Unity Catalog
Creating, populating and using governed feature tables for training and scoring.
MLflow tracking and the model registry
Logging runs, finding the best one, registering models and promoting them with aliases.
Exploring and preparing data
Summary statistics, outliers, visualization, imputation, encoding and transformations on Spark data.
Algorithms, pipelines and imbalance
Choosing algorithms, handling imbalanced data, and building estimator and transformer pipelines.
Hyperparameter tuning and validation
Search strategies, Hyperopt, parallel tuning, and cross-validation versus a single split.
Model evaluation and generalization
Classification and regression metrics, choosing the right one, log-transformed targets, and the bias-variance tradeoff.
Deploying and serving models
Batch, streaming and real-time inference, model endpoints and traffic splitting.
Study all of Machine Learning Associate in the app
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
