AWS Certified Machine Learning Engineer Associate Exam Prep Add-On (MLA-C02 beta; MLA-C01 ended in English on September 28, 2026)
AWS Certified Machine Learning Engineer Associate (MLA-C02) validates the ability to build, operationalize, deploy, and maintain ML and AI solutions on AWS, covering both traditional ML and foundation models. Four domains: Data Preparation for ML and AI, ML Model and Foundation Model (FM) Development, Deployment and Orchestration of ML and AI Workflows, and Operating, Monitoring, and Securing ML and AI Solutions. 65 questions (50 scored, 15 unscored), scaled score 100-1,000, passing score 720. MLA-C02 is in beta as of this writing: beta registration opened September 1, 2026, beta delivery began September 29, 2026, and AWS has not announced a GA date. MLA-C01 remains the live, scorable version through September 28, 2026 in English.
We could not confirm which languages this exam is offered in. Check the owner's page before you book.
Owner: AWS · Read: · Source: AWS Certified Machine Learning Engineer Associate
Exam fee at last check: not confirmed (the owner page did not state a price when we read it), per the vendor. The official objectives are at docs.aws.amazon.com/aws-certification/latest/machine-learning-engineer-associate-02/machine-learning-engineer-associate-02.html. Verified 2026-10-25.
Practice is free. The full analysis adds the right answer, the explanation and the source for every question, readiness by domain and a study plan, for one payment.
- Data Preparation for ML and AI28%
- ML Model and Foundation Model (FM) Development24%
- Deployment and Orchestration of ML and AI Workflows24%
- Operating, Monitoring, and Securing ML and AI Solutions24%
Source: Vendor exam objectives. Checked 2026-10-25. Bar length is relative to the heaviest domain; the figure beside each name is the published weight.
What's on the exam
Data Preparation for ML and AI 28%
The largest domain: collecting and storing data, transforming and engineering features, and validating data quality and bias, across both traditional ML and generative AI pipelines.
- Extracting data from S3, EBS, EFS, RDS, DynamoDB, and OpenSearch Service, and ingesting streaming data with Kinesis, Flink, or Kafka
- Choosing data formats (Parquet, JSON, CSV, ORC) and merging data from multiple sources with Glue or Spark
- Configuring vector databases for AI applications (OpenSearch Service, RDS with pgvector, S3) and ingesting into SageMaker Feature Store
- Feature engineering (scaling, standardization, binning, log transformation) and embedding models that turn text or image data into numerical representations
- Preparing documents for RAG (chunking strategies, metadata extraction) and preparing data for FM fine-tuning, continuous pre-training, and distillation
- Validating data quality, labeling and annotating data, mitigating bias (dataset splitting, shuffling, augmentation), and resolving class imbalance across numeric, text, and image data
Primary sources: AWS MLA-C02 Exam Guide, Content Domain 1
ML Model and Foundation Model (FM) Development 24%
Choosing modeling approaches, training and fine-tuning models, and evaluating ML and GenAI system performance.
- Evaluating and selecting FMs from Amazon Bedrock, and comparing ML models, GenAI models, and algorithms against business requirements (interpretability, latency, cost)
- Selecting RAG architecture patterns, and choosing between custom solutions, managed services, pre-trained models, and FMs
- SageMaker AI built-in algorithms and script mode, hyperparameter optimization (automatic model tuning), and training-time reduction techniques (early stopping, distributed training)
- Preventing overfitting, underfitting, and catastrophic forgetting, and applying AI customization techniques such as task-specific prompt engineering and fine-tuning
- Reproducible experiments with MLflow on SageMaker AI and Bedrock evaluations, model performance baselines, drift detection, and shadow-versus-production variant comparison
- NLP evaluation metrics (BLEU, ROUGE, BERTScore, semantic similarity), human-in-the-loop evaluation frameworks, LLM-as-a-judge, and RAG retrieval accuracy monitoring
Primary sources: AWS MLA-C02 Exam Guide, Content Domain 2
Deployment and Orchestration of ML and AI Workflows 24%
Managing deployment infrastructure, provisioning resources, and building CI/CD pipelines for MLOps and agentic AI workloads.
- Selecting compute environments, deployment orchestrators, and inference strategies (real-time versus batch), and deploying models built outside AWS via SageMaker AI or Bedrock Custom Model Import
- Deploying and configuring agents, including agent communication protocols and agent state management systems
- Provisioning on-demand versus provisioned resources for cost and performance, configuring SageMaker AI endpoints within a VPC, and AI-specific GPU scaling
- Building Bedrock knowledge bases with vector database configuration, document indexing, and retrieval optimization
- CI/CD with CodeBuild, CodeCommit, CodeDeploy, CodePipeline, and CodeConnections, plus automated rollback strategies and automated testing
- Managing model versions with SageMaker Model Registry and MLflow, managing prompts with Bedrock Prompt Management, and orchestrating RAG knowledge base refresh cycles
Primary sources: AWS MLA-C02 Exam Guide, Content Domain 3
Operating, Monitoring, and Securing ML and AI Solutions 24%
Monitoring inference and performance in production, optimizing cost, and securing ML and AI workloads and endpoints.
- Monitoring model performance with CloudWatch generative AI observability, Bedrock Model Evaluation, and drift detection pipelines, plus A/B testing in production
- Monitoring agent performance and coordination (coordination failures, truncated streaming, tool failures)
- Selecting inference instance families, setting up dashboards, and optimizing capacity for cost, performance, and reliability
- Managing FM inference costs and monitoring AI-specific cost patterns (token usage, embedding computation, vector database storage)
- Least-privilege IAM policies for ML and AI systems, VPCs and security groups to isolate workloads, and CI/CD pipeline security scanning (CodeGuru, Inspector)
- Comprehensive monitoring, auditing, and compliance logging (CloudTrail, Config), and safeguards such as Bedrock Guardrails for sensitive data and responsible AI
Primary sources: AWS MLA-C02 Exam Guide, Content Domain 4
Exam-day strategy
Domain 1 (Data Preparation) is 28% of the exam, the single largest domain. Know vector database options (OpenSearch Service, RDS with pgvector, S3), chunking for RAG, and class-imbalance fixes cold.
The other three domains are each 24%, roughly even weight. Do not over-invest study time in Domain 1 at the expense of deployment, orchestration, and security content.
Watch for distractors that confuse batch transform (large offline jobs), real-time endpoints (low-latency online serving), serverless inference (intermittent traffic), and asynchronous inference (large payloads or long processing times).
Cost and security skills recur across domains 3 and 4: expect scenario questions that combine a deployment choice with its cost or security consequence, not just an isolated fact.
MLA-C02 adds explicit agentic-workflow skills not present on MLA-C01 (agent deployment, agent state management, agent communication protocols, coordination-failure monitoring). Do not study from MLA-C01 material alone.
This exam is currently in beta with no announced GA date. If you sit the beta, your score is reported later once AWS calibrates the beta results; the pass/fail designation does not apply until then.
50 of the 65 questions are scored; unscored questions are not identified, so treat every question as if it counts.
How to prepare for AWS Certified Machine Learning Engineer Associate Exam Prep Add-On (MLA-C02 beta; MLA-C01 ended in English on September 28, 2026)
- EVIDENCE-BACKEDPractice in the same format as the exam. How well practice matches the final test shapes how much it helps. (Yang et al., 2021)
- EVIDENCE-BACKEDStart practice questions in week one, even on topics you have not studied. Failed attempts before seeing the answer still helped learning. (Kornell et al., 2009 and others)
- EVIDENCE-BACKEDSpace your reviews over weeks instead of cramming the week before. (Cepeda et al., 2006 and others)
- PLAUSIBLEDivide the minutes by the questions before you start, and keep a margin for a second pass if the exam allows one. (Sound arithmetic; no study compared time budgets on a licensing exam.)
Why this works
EVIDENCE-BACKED At least one peer-reviewed study or official source found the effect or states the fact. A study can be small.
PLAUSIBLE Reasonable and partly studied, but not tested on exams, or the results are mixed.
Practice that matches the final test format works better than practice that does not.
Failed attempts before seeing the answer still improved later learning of the material asked about; the benefit for material not asked about was virtually nil.
Spacing beat massing across hundreds of experiments and beat cramming with flashcards; combined with retrieval it raised scores on a real high-stakes exam.
Sound arithmetic; no study compared time budgets on a licensing exam.
Sources
- Yang, C., Luo, L., Vadillo, M. A., Yu, R., & Shanks, D. R. (2021). Testing (quizzing) boosts classroom learning: A systematic and meta-analytic review. Psychological Bulletin, 147(4), 399-435. doi.org/10.1037/bul0000309
- Kornell, N., Hays, M. J., & Bjork, R. A. (2009). Unsuccessful retrieval attempts enhance subsequent learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(4), 989-998. doi.org/10.1037/a0015729
- Richland, L. E., Kornell, N., & Kao, L. S. (2009). The pretesting effect: Do unsuccessful retrieval attempts enhance learning? Journal of Experimental Psychology: Applied, 15(3), 243-257. doi.org/10.1037/a0016496
- St. Hilaire, K. J., Chan, J. C. K., & Ahn, D. (2024). Guessing as a learning intervention: A meta-analytic review of the prequestion effect. Psychonomic Bulletin & Review, 31(2), 411-441. doi.org/10.3758/s13423-023-02353-8
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354-380. doi.org/10.1037/0033-2909.132.3.354
- Kornell, N. (2009). Optimising learning using flashcards: Spacing is more effective than cramming. Applied Cognitive Psychology, 23(9), 1297-1317. doi.org/10.1002/acp.1537
- Kang, S. H. K. (2016). Spaced repetition promotes efficient and effective learning: Policy implications for instruction. Policy Insights from the Behavioral and Brain Sciences, 3(1), 12-19. doi.org/10.1177/2372732215624708
- Janes, J. L., Dunlosky, J., Rawson, K. A., & Jasnow, A. (2020). Successive relearning improves performance on a high-stakes exam in a difficult biopsychology course. Applied Cognitive Psychology, 34(5), 1118-1132. doi.org/10.1002/acp.3699
Additional resources
Explore Related Cybersecurity Resources
DecipherU is independent of Amazon Web Services. Certified Machine Learning Engineer Associate (MLA-C01 current; MLA-C02 exam guide published, beta only, GA date TBD) is a trademark of Amazon Web Services. DecipherU is not endorsed by, sponsored by, or affiliated with any certifying body.