AWS Machine Learning Study Roadmap

Domain 1: Data Repositories — S3 data lakes, Redshift, RDS, DynamoDB for ML pipelines

Domain 1: Data Ingestion — Kinesis Data Streams/Firehose, Glue ETL, Data Pipeline, batch vs streaming

Domain 1: Data Transformation — Glue DataBrew, EMR Spark, Athena, feature engineering techniques

Domain 2: Data Sanitization — Missing values, outliers, normalization, encoding categorical variables

Domain 2: Feature Engineering — Feature selection, dimensionality reduction (PCA), data visualization

Domain 2: Data Analysis — Statistical analysis, correlation, distribution analysis, SageMaker Data Wrangler

Domain 3: ML Frameworks — SageMaker built-in algorithms, XGBoost, Linear Learner, image classification

Domain 3: Model Training — Hyperparameter tuning, training jobs, distributed training, Spot instances

Domain 3: Evaluation — Confusion matrix, precision/recall, F1 score, AUC-ROC, overfitting/underfitting

Domain 4: SageMaker Endpoints — Real-time inference, batch transform, multi-model endpoints, auto-scaling

Domain 4: MLOps — SageMaker Pipelines, Model Registry, Model Monitor, A/B testing, CI/CD for ML

Domain 4: Security & Cost — VPC configs, KMS encryption, IAM roles, Spot training, instance selection

Full Review: Practice exams, SageMaker hands-on labs, Algorithm review, Exam logistics

Final Review: Weak areas deep-dive, Full-length practice tests, Time management strategies