Directory / Amazon Web Services

AWS Certified Generative AI Developer – Professional

The AWS credential aimed squarely at developers building LLM applications in production: RAG, vector stores, prompt management, agents, guardrails and evaluation on Amazon Bedrock.

What it signals

Practical assessment: Scenario-based multiple choice. Proctored scenario-based multiple choice/multiple response only.
Identity assurance: Proctored with identity verification. Pearson VUE testing centre or online proctored
Depth: Professional. Professional level.
Experience expected: 2 to 3 years. Guide recommends 2+ years building production applications plus 1 year of GenAI.
Portability: Mixed. Patterns (RAG, agents, MCP, evaluation) are transferable, but solutions are expressed in AWS services.
Verifiability: Public badge or registry. Credly digital badge with verification.

For hiring managers

The most relevant AWS credential for developers building LLM applications (RAG, vector stores, prompt management, agents, guardrails, evaluation).
New: beta ran Oct 2025 - Mar 31 2026, so every holder earned it recently; small holder population.
Despite the 'Professional' label and scenario depth, it is still a selected-response exam - no build task. Pair with the hands-on microcredential or a portfolio check.
Not an ML modelling credential: model training is explicitly out of scope.
Blueprint names fast-moving products (Strands Agents, Agent Squad, AgentCore, MCP); specifics may date quickly.

Verifying it

Credly digital badge link shared by the candidate. No public registry. Beta passers may hold an additional 'Early Adopter' badge (not verified on an official page for this exam).

What it covers

Published exam outline

Foundation Model Integration, Data Management, and Compliance 31%
Implementation and Integration 26%
AI Safety, Security, and Governance 20%
Operational Efficiency and Optimization for GenAI Applications 12%
Testing, Validation, and Troubleshooting 11%

By topic

GenAI apps: Core focus (over 35%). The whole blueprint is about building on foundation models; Domains 1, 2 and 5 alone (FM integration, implementation, evaluation) are 68%.
Agents: Some (10 to 20%). Agentic AI is one of five tasks in Domain 2 (26%) plus scattered skills; estimated 10-15%, no official sub-weight.
Classical ML: Not covered. Model development and training are explicitly out of scope.
Operations: Substantial (20 to 35%). Deployment strategies and CI/CD in Domain 2, Domain 4 operations (12%) and Domain 5 troubleshooting (11%) total roughly 25-30%.
Data: Some (10 to 20%). Domain 1 includes data management, vector stores and retrieval data pipelines, but data engineering is out of scope; estimated 10-15%.
Responsible AI: Substantial (20 to 35%). Domain 3 AI Safety, Security, and Governance is 20%, plus compliance in Domain 1.

Exam details

Who the provider says it is for. Individuals who perform a GenAI developer role: integrating foundation models into applications and business workflows and implementing GenAI solutions in production on AWS.

Format: multiple choice; multiple response
Delivery: Pearson VUE testing centre or online proctored
Code or hands-on work: No labs and no code writing; multiple choice and multiple response only. Model training, advanced ML and data/feature engineering are out of scope. Whether code snippets must be read is not stated in the guide.
Prerequisites: None required. AWS suggests (optional) AI Practitioner, Solutions Architect - Associate, ML Engineer - Associate or Data Engineer - Associate first.
Recommended experience: 2 or more years building production-grade applications on AWS or with open-source technologies, general AI/ML or data engineering experience, and 1 year of hands-on experience implementing GenAI solutions.
Renewal: Pass the latest version of the exam (50% voucher) or Skill Builder maintenance option. Passing AIP-C01 also renews AI Practitioner, ML Engineer - Associate and Data Engineer - Associate.

Preparing

Not confirmed

These details were not available from an official source at the time of research.

Exact general-availability date (2026-03-17 came from a search summary, not an official page)
Beta price and beta question count
Exact share of blueprint for agents, data and mlops (estimates)
Whether code snippets appear in exam items

Sources (5)