AWS Machine Learning Specialty (MLS-C01) Study Plan¶

14-Week Comprehensive Study Schedule¶

Prerequisites Check¶

Python programming proficiency (pandas, numpy, scikit-learn)
Statistics and linear algebra fundamentals
AWS Associate-level knowledge (Solutions Architect or Developer recommended)
Basic machine learning concepts understanding

Week 1-3: Data Engineering for ML (20% of exam)¶

Week 1: Data Ingestion and Storage¶

S3 for data lakes and lifecycle policies
AWS Glue Data Catalog and crawlers
Kinesis Data Streams and Firehose for real-time data
Data formats: Parquet, ORC, Avro, JSON
Hands-on: Build data lake on S3 with Glue catalog
Lab: Real-time data pipeline with Kinesis
Review Notes: 01-data-engineering.md

Week 2: Data Transformation¶

AWS Glue ETL jobs and PySpark
AWS Glue DataBrew for visual data prep
Amazon EMR for big data processing
Data partitioning and optimization strategies
Hands-on: Create ETL pipeline with Glue
Lab: Process large datasets with EMR Spark

Week 3: Feature Store and Data Quality¶

SageMaker Feature Store implementation
Data versioning and lineage
Data quality validation and monitoring
Lake Formation for data access control
Hands-on: Implement feature store for ML features
Lab: Data quality pipeline with automated validation

Week 4-6: Exploratory Data Analysis (24% of exam)¶

Week 4: Data Preprocessing¶

Handling missing values and outliers
Data normalization and standardization
Categorical encoding techniques
Feature scaling and transformation
Hands-on: Comprehensive data cleaning pipeline
Lab: SageMaker Data Wrangler for EDA
Review Notes: 02-exploratory-data-analysis.md

Week 5: Feature Engineering¶

Feature selection methods (filter, wrapper, embedded)
Feature extraction and dimensionality reduction (PCA, t-SNE)
Time series feature engineering
Text feature engineering (TF-IDF, word embeddings)
Image feature engineering (CNNs, transfer learning)
Hands-on: Build feature engineering pipeline
Lab: Advanced feature selection with scikit-learn

Week 6: Data Visualization and Analysis¶

Amazon QuickSight for ML data visualization
Statistical analysis and hypothesis testing
Correlation analysis and multicollinearity
Data distribution analysis
Bias detection in datasets
Hands-on: Interactive dashboards with QuickSight
Lab: Statistical analysis with SageMaker notebooks

Week 7-10: Modeling (36% of exam - largest domain)¶

Week 7: ML Problem Formulation and Algorithm Selection¶

Supervised learning: regression, classification
Unsupervised learning: clustering, dimensionality reduction
Reinforcement learning fundamentals
SageMaker built-in algorithms overview
Hands-on: Problem formulation for business scenarios
Lab: Compare multiple algorithms for same problem
Review Notes: 03-modeling.md

Week 8: SageMaker Training¶

SageMaker training jobs and instance types
Distributed training strategies
Spot training for cost optimization
Custom algorithms with containers
Hands-on: Train models with SageMaker built-in algorithms
Lab: Custom algorithm deployment with Docker

Week 9: Hyperparameter Tuning¶

SageMaker Automatic Model Tuning
Hyperparameter search strategies (Bayesian, random, grid)
Early stopping for efficient tuning
Multi-objective hyperparameter optimization
Hands-on: Optimize model with automatic tuning
Lab: Compare tuning strategies

Week 10: Model Evaluation¶

Evaluation metrics: accuracy, precision, recall, F1, AUC-ROC
Cross-validation techniques
Confusion matrix analysis
Bias-variance tradeoff
Model interpretability and explainability
Hands-on: Comprehensive model evaluation pipeline
Lab: SageMaker Clarify for bias detection

Week 11-12: ML Implementation and Operations (20% of exam)¶

Week 11: Model Deployment¶

SageMaker real-time endpoints
Batch transform jobs
Multi-model endpoints
Inference pipelines
A/B testing and canary deployments
Hands-on: Deploy model to production endpoint
Lab: Implement A/B testing for model versions
Review Notes: 04-ml-implementation-operations.md

Week 12: MLOps and Monitoring¶

SageMaker Model Monitor for data drift
SageMaker Pipelines for ML workflows
Model registry and versioning
CI/CD for ML with CodePipeline
Auto-scaling endpoints
Hands-on: Build end-to-end ML pipeline
Lab: Implement model monitoring and alerting

Week 13: AWS AI Services and Advanced Topics¶

Week 13: AI Services Deep Dive¶

Amazon Rekognition for computer vision
Amazon Comprehend for NLP
Amazon Textract for document processing
Amazon Transcribe for speech-to-text
Amazon Translate for language translation
Amazon Forecast for time series
Amazon Personalize for recommendations
Hands-on: Build applications with AI services
Lab: Custom models with Rekognition and Comprehend

Week 14: Final Review and Practice¶

Week 14: Comprehensive Review¶

Review all domain notes and weak areas
Complete full-length practice exam (3 hours)
Analyze results by domain
Hands-on practice with SageMaker
Review ML algorithms and when to use each
Final practice exam (aim for 85%+)
Exam-taking strategies review
Schedule certification exam

Daily Study Routine (3-4 hours/day)¶

Recommended Schedule¶

45 minutes: AWS ML documentation and whitepapers
120 minutes: Hands-on labs with SageMaker and Python
45 minutes: Practice questions and algorithm review
15 minutes: Note-taking and concept reinforcement

Essential Hands-on Projects¶

Project 1: End-to-End ML Pipeline (Weeks 1-6)¶

Data ingestion from multiple sources
ETL pipeline with AWS Glue
Feature engineering and selection
Data quality validation
Feature store implementation
Exploratory data analysis with QuickSight

Project 2: Image Classification System (Weeks 7-9)¶

Custom CNN training with SageMaker
Transfer learning with pre-trained models
Hyperparameter tuning
Model evaluation and comparison
Error analysis and model improvement

Project 3: NLP Sentiment Analysis (Weeks 7-9)¶

Text preprocessing and feature extraction
Multiple algorithm comparison (SageMaker built-ins)
Hyperparameter optimization
Model interpretability
Production deployment

Project 4: Time Series Forecasting (Weeks 7-9)¶

Project 5: Production ML System (Weeks 11-12)¶

Practice Exam Strategy¶

Target Scores by Week¶

Week 6: 60%+ on practice exams
Week 9: 70%+ on practice exams
Week 12: 80%+ on practice exams
Week 14: 85%+ consistently

Comprehensive Study Resources¶

👉 Complete AWS Study Resources Guide

Quick Links (MLS-C01 Specific)¶

MLS-C01 Official Exam Page - Registration and details
AWS Machine Learning Skill Builder - FREE ML courses
SageMaker Documentation - Complete SageMaker guide
AWS ML Blog - Latest ML on AWS

Essential Resources¶

AWS Whitepapers¶

Machine Learning Lens - AWS Well-Architected Framework
Amazon SageMaker Developer Guide
Power Machine Learning at Scale
MLOps Foundations Roadmap

Recommended Courses¶

AWS Skill Builder: Exam Readiness - Machine Learning Specialty
Stephane Maarek: AWS Certified Machine Learning Specialty
A Cloud Guru: Machine Learning Specialty Certification

Practice Exams¶

Tutorials Dojo MLS-C01 Practice Tests ($14.99) - Highly recommended
AWS Official Practice Exam ($40)
Whizlabs ML Specialty Practice Tests

Key Exam Topics Summary¶

Data Engineering (20%)¶

Data lakes and storage optimization
ETL/ELT pipelines with Glue
Real-time data ingestion with Kinesis
Feature stores and data versioning

EDA (24%)¶

Data cleaning and preprocessing
Feature engineering techniques
Statistical analysis
Data visualization

Modeling (36%)¶

Algorithm selection for problem types
SageMaker training and built-in algorithms
Hyperparameter tuning
Model evaluation metrics

ML Implementation (20%)¶

Deployment strategies
Model monitoring and drift detection
MLOps with SageMaker Pipelines
Cost and performance optimization

SageMaker Built-in Algorithms Checklist¶

Supervised Learning¶

Linear Learner (regression/classification)
XGBoost (regression/classification)
Factorization Machines (recommendations)
K-Nearest Neighbors (KNN)
Image Classification (ResNet)
Object Detection (SSD)
Semantic Segmentation
Sequence to Sequence (translation)
BlazingText (text classification, word2vec)

Unsupervised Learning¶

K-Means (clustering)
Principal Component Analysis (PCA)
Random Cut Forest (anomaly detection)
IP Insights (fraud detection)
Neural Topic Model (topic modeling)
Latent Dirichlet Allocation (LDA)

Time Series¶

DeepAR Forecasting

AWS AI Services Proficiency¶

Final Exam Checklist¶

Technical Preparation¶

Hands-on with all SageMaker features
Built at least 3 end-to-end ML projects
Understand all built-in algorithms
Proficient with Python ML libraries
Know when to use which AI service

Exam Day Strategy¶

Time management: ~2.8 minutes per question
Read scenarios carefully for problem type
Identify constraints (cost, latency, accuracy)
Eliminate obviously incorrect algorithms
Consider scalability and operational requirements
Flag uncertain questions for review

Common Exam Traps¶

Choosing complex solutions when simple ones suffice
Ignoring cost optimization opportunities
Not considering data preprocessing requirements
Overlooking model monitoring in production
Missing hyperparameter tuning opportunities
Forgetting about data security and privacy

Post-Certification Path¶

Advanced ML specializations (deep learning, reinforcement learning)
MLOps certifications
Contribute to ML open-source projects
Write blog posts about ML implementations
Participate in Kaggle competitions
Mentor others in ML journey