What this credential is — and the C01 → C02 change

Lesson 1 of 4 in The MLA Exam: Blueprint, Logistics, Strategy.

The AWS Certified Machine Learning Engineer – Associate validates that you can build, deploy, and operate machine learning workloads on AWS — not that you can derive backpropagation on a whiteboard. It sits at the associate level: deeper than AI Practitioner, broader and more operational than the old Machine Learning – Specialty (which retired on March 31, 2026 and points its former audience here).

AWS recommends at least a year of hands-on work with Amazon SageMaker AI, Amazon Bedrock, and the surrounding data and DevOps services, in a role like ML engineer, MLOps engineer, LLMOps engineer, data engineer, or data scientist. The credential is valid for three years.

How the credential got here

  • 2024-07-22 — AWS announces new exam question types:

    Ordering, matching, and case-study items join multiple choice/response — with AI Practitioner and the new ML Engineer – Associate named as the first exams to carry them.

  • 2024-07-31 — MLA-C01 exam guide v1.0 published:

    The first exam guide for the ML Engineer – Associate: 65 questions (50 scored + 15 unscored), 130 minutes, scaled 100–1,000 with pass at 720, domains weighted 28/26/22/24.

  • 2026-03-31 — ML – Specialty (MLS-C01) retires:

    The last day to take the harder, older Machine Learning – Specialty. AWS points ML candidates to the ML Engineer – Associate as its ongoing machine-learning credential.

  • 2026-07-14 — AWS announces the MLA-C02 revision:

    The update brings foundation models, RAG, and agents into the blueprint and names LLMOps engineers among the target roles.

  • 2026-09-01 — MLA-C02 exam guide released; beta registration opens:

    The C02 guide lands on docs.aws.amazon.com with rebalanced domains (28/24/24/24), multiple choice + multiple response only, and Bedrock/AgentCore on the in-scope list.

  • 2026-09-28 — Last day for MLA-C01 in English:

    English sittings end; Japanese, Korean, and Simplified Chinese continue through the beta period until C02 reaches standard delivery.

  • 2026-09-29 — MLA-C02 beta delivery begins:

    The beta form: 85 questions, 170 minutes, $75, English only — full three-year credential for passers, and results within 5 business days like any other exam (AWS states this explicitly "including results for beta exams"; the "beta scores take months" folklore is wrong).

  • 2025-10-14 — AWS announces the Generative AI Developer – Professional cert:

    A new Professional-level credential (AIP-C01) for building production GenAI on AWS opens in beta, announced alongside the Agentic AI Demonstrated microcredential.

  • 2026-03-17 — AIP-C01 standard exam opens — refreshed with AgentCore:

    Registration opens for the standard version, refreshed to cover Amazon Bedrock AgentCore. 75 questions (65 scored + 10 unscored), scaled 100–1,000, pass 750, 180 minutes.

  • 2026-03-31 — Last day for the AIP-C01 beta:

    The beta window closes; the refreshed standard exam carries on as the live form.

  • 2027-01-14 — MLA-C02 standard delivery (announced):

    AWS's September 2026 blog gives January 14, 2027 for GA — the certification page still said "TBD" in October 2026, so verify before booking. MLA-C01 retires everywhere at GA.

What changed under the hood, per AWS's own comparison page: Domain 2 is now "ML Model and Foundation Model Development", every domain gained GenAI task statements (RAG document preparation, FM fine-tuning data, agent deployment, token-cost optimization, Bedrock Guardrails), and a few C01 topics were dropped — model-size reduction with pruning/compression, SageMaker Neo edge optimization, and bring-your-own-container depth. The four domains rebalanced from 28/26/22/24 to 28/24/24/24.

Key terms: MLA-C02, MLA-C01, Amazon SageMaker AI, Amazon Bedrock

Interactive checkpoint quiz (2 questions) — open this page in a browser to take it.