When you pass each exam, you'll be certification ready.
0%
Pass Percentage
0
Five Star Reviews
0
Certifications
0
Cups of coffee
AWS Certified Machine Learning Specialty Exam Topics
Exam Basics
- Format: 50 scored questions and 15 unscored questions
- Question types: multiple choice and multiple response
- Exam time: about 180 minutes
- Passing score: 750 out of 1000
- Target candidate: 2+ years of ML or deep learning on AWS
- Scaled scoring ensures fairness across exam versions
Domain 1: Data Engineering (20%)
- Create data repositories for ML with S3, EFS, and EBS
- Implement data ingestion solutions using Glue, Kinesis, EMR, and Flink
- Schedule and orchestrate batch and streaming pipelines
- Transform and preprocess data with ETL and MapReduce tools
Domain 2: Exploratory Data Analysis (24%)
- Sanitize and prepare datasets: handle missing and corrupt data
- Perform feature engineering from text, speech, image, and public datasets
- Apply binning, tokenization, one-hot encoding, and dimensionality reduction
- Visualize and interpret data with scatter plots, histograms, and clustering
Domain 3: Modeling (36%)
- Frame business problems as ML problems
- Select models such as regression, XGBoost, decision trees, CNNs, RNNs, and LLMs
- Train ML models with cross-validation, gradient descent, GPUs, and distributed training
- Perform hyperparameter optimization with dropout, L1/L2, and learning rate tuning
- Evaluate models using accuracy, precision, recall, F1, RMSE, ROC/AUC, and confusion matrices
- Conduct A/B testing and cross-validation to compare models
Domain 4: Implementation & Operations (20%)
- Deploy scalable, resilient, and fault-tolerant ML solutions
- Use CloudTrail, CloudWatch, Auto Scaling, and multi-Region deployments
- Choose ML services such as SageMaker, Polly, Lex, Transcribe, and Q
- Apply AWS security practices with IAM, S3 policies, VPCs, and encryption
- Expose endpoints, retrain pipelines, and monitor model performance
Out of Scope
- Extensive algorithm development and complex mathematical proofs
- Advanced hyperparameter optimization
- Deep networking, security, or DevOps design
- DevOps tasks for EMR and unrelated AWS services
- Architecture and troubleshooting outside ML scope
How to Prepare
- Study the official exam guide and review domain task statements
- Practice data engineering, feature engineering, and modeling in AWS
- Use SageMaker and AWS ML services in hands-on projects
- Take practice exams to get familiar with multiple-choice and multiple-response formats
- Review unscored practice questions as learning opportunities
Certification Made Easy
Get certified with the help of Cameron McKenzie (@cameronmcnz), the industry's foremost expert AI, ML, DevOps and Cloud Computing.
- Practical exam-style questions with clear explanations
- Short lessons that fit your schedule
- Targeted study roadmaps for each certification
- Expert tips on applying knowledge in real projects
- Active community support and study groups
- Fast refreshers and final-day checklists