AWS · MLA-C01
AWS Certified Machine Learning Engineer - Associate (MLA-C01) Practice Tests
Validates ability to build, operationalize, deploy, and maintain machine learning solutions and pipelines using AWS Cloud services.
Exam Details
Practice Questions
582
≈ 8 practice exams
Duration
130 minutes
Passing Score
720/1000
Difficulty
AssociateLast Updated
Oct 2026
Topics Covered
Data PreparationML Model DevelopmentDeployment and OrchestrationMonitoring, Maintenance, and SecurityAmazon SageMaker
AWS MLA-C01 Machine Learning Engineer Practice Exam
MLA-C01 validates production machine learning work across data preparation (28%), model development (26%), deployment and orchestration (22%), and monitoring, maintenance, and security (24%). It focuses on building and operating ML workloads with services such as Amazon SageMaker, AWS Glue, Amazon S3, Step Functions, IAM, and CloudWatch rather than on algorithm theory alone.
The exam has 65 questions in 130 minutes, costs $150 USD, and uses a 100-1,000 scale with 720 required to pass. AWS recommends at least one year using SageMaker and related ML engineering services. A credential earned from MLA-C01 remains valid for three years.
MLA-C01 is now a legacy version: English testing ended September 28, 2026, while Japanese, Korean, and Simplified Chinese remain available only until MLA-C02 reaches general availability. Use these 582 questions only when your booking is specifically MLA-C01; candidates booking the new English exam should use the separate MLA-C02 page because its beta adds generative AI and agentic-workflow coverage.
Exam Domain Breakdown
Data Preparation for Machine Learning28%
ML Model Development26%
Deployment and Orchestration of ML Workflows22%
ML Solution Monitoring, Maintenance, and Security24%
Exam Overview
AWS Certified Machine Learning Engineer – Associate (MLA-C01) validates the ability to build, operationalize, deploy, and maintain machine learning solutions and pipelines on AWS. Its four domains cover data preparation, model development, deployment and orchestration, and monitoring, maintenance, and security.
MLA-C01 is now a legacy exam version. English testing ended September 28, 2026; AWS says Japanese, Korean, and Simplified Chinese remain available until MLA-C02 reaches general availability. Credentials already earned through MLA-C01 keep their normal three-year validity.
Who Should Take This Exam
MLA-C01 is intended for machine learning engineers and adjacent backend, DevOps, and data engineering professionals who operationalize ML workloads on AWS. AWS recommends at least one year using Amazon SageMaker and related ML engineering services.
Candidates booking in English should prepare for MLA-C02 instead. This page remains useful for candidates whose confirmed appointment is specifically MLA-C01 in one of the remaining languages and for holders reviewing the legacy outline.
Prerequisites
AWS has no mandatory certification prerequisite for MLA-C01. Its target candidate has at least one year using SageMaker and other AWS ML engineering services, plus experience in a related role such as backend development, DevOps, data engineering, or data science.
Always match preparation to the code shown in the booking system: MLA-C02 expands the blueprint, so MLA-C01 and MLA-C02 study materials are not interchangeable.
Exam Format
MLA-C01 contains 65 questions and allows 130 minutes. AWS uses multiple-choice, multiple-response, ordering, matching, and case-study formats, scores the exam from 100 to 1,000, and requires 720 to pass. The listed price is $150 USD, and delivery is through Pearson VUE.
English registration and testing closed September 28, 2026. AWS says Japanese, Korean, and Simplified Chinese remain available only until MLA-C02 reaches general availability. A certification earned through MLA-C01 remains valid for three years.
Skills Measured
- 1.Domain 1: Data Preparation for Machine Learning (28%) — Ingesting data from AWS sources, transforming and validating datasets, feature engineering, handling class imbalance, splitting datasets, and using Amazon S3, AWS Glue, Amazon Athena, and SageMaker Data Wrangler.
- 2.Domain 2: ML Model Development (26%) — Selecting ML algorithms and frameworks, training models with Amazon SageMaker, tuning hyperparameters via SageMaker Automatic Model Tuning, analyzing performance metrics, managing versions with the SageMaker Model Registry, and applying responsible AI practices.
- 3.Domain 3: Deployment and Orchestration of ML Workflows (22%) — Selecting deployment infrastructure and endpoint types (real-time, batch, asynchronous), provisioning compute resources, configuring auto scaling, and building CI/CD pipelines with AWS CodePipeline, AWS Step Functions, and Amazon SageMaker Pipelines.
- 4.Domain 4: ML Solution Monitoring, Maintenance, and Security (24%) — Monitoring model performance and data drift with SageMaker Model Monitor, logging and alerting via Amazon CloudWatch, retraining triggers, IAM policies, VPC configuration, encryption, and compliance/auditability.
Study Tips
- Confirm that your booking says MLA-C01 before using this outline; new English appointments use MLA-C02.
- Map study sessions to the official weights: Data Preparation 28%, Model Development 26%, Deployment and Orchestration 22%, and Monitoring, Maintenance, and Security 24%.
- Practise a complete SageMaker workflow: prepare data, train and tune, register a model, deploy it, monitor it, and trigger maintenance or retraining.
- Study service choices in context, including S3, Glue, SageMaker, Step Functions, CodePipeline, CloudWatch, IAM, networking, and encryption.
- Rehearse ordering, matching, and case-study items as well as ordinary multiple-choice questions.
- Use the separate MLA-C02 materials if your objective includes the new generative AI and agentic-workflow coverage.
Career Benefits
MLA-C01 demonstrates associate-level AWS machine learning engineering skills, especially the production work between a model prototype and a monitored service. It is relevant to ML engineer, MLOps, data engineering, and platform roles that use AWS.
Because the exam version is retiring, the durable value is the active three-year AWS certification and the underlying operational skill. Candidates beginning now should normally follow the MLA-C02 path, while existing MLA-C01 holders can continue to present the credential until its individual expiration date.
Sample Questions
5 sample questions with answers and explanations. The full bank has 582 questions, enough for 8 full-length practice exams.
Preview — answers shown1. When ingesting data into the SageMaker Feature Store Online Store, which API call is used?
A `PutRecord`
B `CreateFeatureGroup`
C `BatchGetRecord`
D `GetRecord`
Explanation
The correct API call is `PutRecord`. Here's why: The `PutRecord` API is the primary method for ingesting or updating feature values in the SageMaker Feature Store. You provide the name of the Feature Group and a single record containing the feature values. When you call `PutRecord`, SageMaker updates the feature values in both the low-latency Online Store (if enabled) and appends the new values to the historical Offline Store (if enabled). Why the other options are incorrect: - A & C: `GetRecord` and `BatchGetRecord` are used to *retrieve* data from the Online Store, not to ingest it. - B: `CreateFeatureGroup` is a one-time setup operation to define the feature group's schema; it does not handle data ingestion.
2. Which SageMaker deployment strategy is MOST analogous to a traditional database 'batch job'?
AAsynchronous Inference
BServerless Inference
CReal-time Inference
DBatch Transform
Explanation
Batch Transform is the most analogous to a traditional batch job. Here's why: Like a traditional database batch job, a SageMaker Batch Transform job is designed to run on a schedule or be initiated once to process a large, predefined set of data. It is an offline process that runs for a finite duration, processes the entire input, writes the output, and then shuts down. This contrasts with the other options, which are all designed as persistent, on-demand services for handling individual requests. Why the other options are incorrect: - A, B, C: These are all services designed to be 'always on' to serve individual, on-demand requests, which is the opposite of a batch job paradigm.
3. Which of the following metrics would be used to evaluate a regression model, NOT a classification model?
AF1-Score
BMAE (Mean Absolute Error)
CAUC (Area Under the ROC Curve)
DPrecision
Explanation
MAE is a regression metric. Here's why: Evaluation metrics are specific to the type of ML problem. **Regression** problems (predicting a continuous value like price or temperature) are evaluated based on the magnitude of the prediction error. Metrics like Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R-squared are used for this. **Classification** problems (predicting a discrete category like 'spam' or 'not spam') are evaluated based on how many predictions are correct. Metrics like Precision, Recall, F1-Score, and AUC are used for this. Why the other options are incorrect: - A, B, C: These are all standard and widely used classification metrics.
4. A hyperparameter tuning job for a deep learning model is configured to monitor 'validation_loss'. What `objective_type` should be set for the tuner?
AMaximize
BMinimize
CEqualize
DCategorical
Explanation
The objective type should be 'Minimize'. Here's why: The goal of model training is typically to minimize the loss function. A lower loss (or error) indicates that the model's predictions are closer to the actual ground truth values. Therefore, when you are tuning hyperparameters to find the best model, you want to find the combination that results in the lowest possible value for 'validation_loss'. This corresponds to an `objective_type` of 'Minimize'. Why the other options are incorrect: - A: 'Maximize' would be used for metrics where higher is better, such as 'accuracy' or 'AUC'. - C: 'Equalize' is not a valid objective type. - D: 'Categorical' is a type of hyperparameter range, not an objective type.
5. For a binary classification model, which metric represents the proportion of true positives out of all the instances that the model predicted as positive?
ARecall
BAccuracy
CF1-Score
DPrecision
Explanation
This metric is Precision. Here's why: Precision answers the question: 'Of all the predictions I made for the positive class, how many were actually correct?'. It is calculated as `True Positives / (True Positives + False Positives)`. It is a critical metric when the cost of a false positive is high. For example, in an email spam filter, high precision is needed to ensure that important emails (false positives) are not sent to the spam folder. Why the other options are incorrect: - A: Accuracy is the overall proportion of correct predictions. - B: Recall is the proportion of actual positives that were correctly identified. - C: F1-Score is the harmonic mean of Precision and Recall.
Frequently Asked Questions
Can I still take MLA-C01?
English testing ended September 28, 2026. AWS says Japanese, Korean, and Simplified Chinese remain available until MLA-C02 reaches general availability; verify the version and language in the booking system.
How many questions are on MLA-C01?
MLA-C01 has 65 questions and allows 130 minutes.
What score passes MLA-C01?
AWS uses a 100-1,000 scale and requires 720 to pass. The exam uses compensatory scoring across the full test.
What are the MLA-C01 domain weights?
Data Preparation is 28%, Model Development 26%, Deployment and Orchestration 22%, and Monitoring, Maintenance, and Security 24%.
Does MLA-C01 require another AWS certification?
No. AWS recommends one year using SageMaker and related ML engineering services, but does not require a prior certification.
Does an MLA-C01 credential expire because the exam retired?
No. A credential earned by passing MLA-C01 remains valid for its normal three-year term.
More AWS Practice Exams
AWS Certified Machine Learning Engineer - Associate (MLA-C02)
MLA-C02 · 321 questions
AWS Certified Machine Learning - Specialty (MLS-C01)
MLS-C01 · 860 questions
AWS Certified Generative AI Developer - Professional (AIP-C01)
AIP-C01 · 1978 questions
AWS Certified Developer - Associate (DVA-C02)
DVA-C02 · 536 questions
AWS Certified DevOps Engineer - Professional (DOP-C02)
DOP-C02 · 436 questions
AWS Certified Security - Specialty (SCS-C03)
SCS-C03 · 2069 questions
$17.99
One-time access to this exam
582 questions (8 practice exams' worth)
Unlimited timed exam simulations
Yours forever — no subscription
Detailed explanations
Free preview stays available