AWS MLA-C02 2026 考試指南：Beta 時程、領域比重與温習計劃

AWS is updating the Certified Machine Learning Engineer – Associate exam from MLA-C01 to MLA-C02. The MLA-C02 beta has been running since 29 September 2026. Here are the key dates, the format, the four domains and their weights, the recommended experience, and a study plan.

The key dates

The 14 January 2027 date comes from AWS’s September 2026 announcement. The exam page still lists general availability as “TBD”, so check it before you plan around that date. Until then, MLA-C01 remains available in Japanese, Korean and Simplified Chinese.

The format

• MLA-C02 beta: 85 questions in 170 minutes, USD 75 (beta pricing), English only
• Question types: multiple choice and multiple response. Unanswered questions are scored as incorrect, and there is no penalty for guessing
• Delivery: Pearson VUE, at a test centre or online proctored
• Scoring (MLA-C02 exam guide): the guide describes 50 scored and 15 unscored questions, a scaled score from 100 to 1,000, and a minimum passing score of 720. It notes that the pass or fail designation does not apply to the beta
• Validity: 3 years. You recertify by passing the latest version of the exam

The exam page does not yet give the duration or price of the standard (non-beta) MLA-C02 exam. For comparison, MLA-C01 has 65 questions in 130 minutes and costs USD 150.

Under AWS’s beta rules, you can take a beta only once, results are available within 5 business days, and a pass earns the certification, active for three years.

The four domains

AWS says the domain structure stays the same and no new domains were added. The names and two of the weights have changed:

The new content is mostly generative and agentic AI: vector databases, embeddings, preparing documents for retrieval augmented generation (RAG), choosing and fine-tuning foundation models in Amazon Bedrock, agents, and Amazon Bedrock Guardrails. AWS’s comparison page also lists removed items, including SageMaker Neo for edge devices and bring your own container (BYOC) with SageMaker.

Recommended experience

The MLA-C02 exam guide describes the target candidate as having:

• at least 1 year of experience using Amazon SageMaker AI, Amazon Bedrock and other AWS services for ML engineering
• at least 1 year in a related role, such as backend software developer, DevOps developer, data engineer or data scientist
• experience with both traditional ML and generative AI

The guide also lists out-of-scope tasks, such as designing full end-to-end AI and ML solutions, and working deeply in two or more ML domains.

A study plan by weight

The weights are close together, so give every domain real time. If you have about 50 hours, a split that follows the weights is:

• 14 hours: Data Preparation for ML and AI (collecting and storing data, transformation and feature engineering, data quality and bias)
• 12 hours: ML Model and FM Development (choosing a modelling approach, training and fine-tuning, evaluation)
• 12 hours: Deployment and Orchestration (deployment infrastructure, provisioning resources, CI/CD pipelines)
• 12 hours: Operating, Monitoring, and Securing (monitoring inference, managing cost and performance, securing endpoints)

If you studied for MLA-C01, spend extra time on the new generative AI skills, because they appear in every domain.

Confirm the latest details on AWS’s Machine Learning Engineer – Associate page before you book.

I run ExamCrest, where this post first appeared. If you're preparing for MLA-C02, there are 10 free MLA-C02 practice questions, each with the answer and a full explanation.