AWS Certified Generative AI Developer – Professional (AIP-C01)
Design, build, tune, evaluate, and deploy production-ready generative AI, autonomous agentic workflows, multi-modal applications, and LLM solutions using Amazon Bedrock and AWS AI services.
Official Exam Specification & Quick Facts
Question Difficulty Breakdown
Recommended Study Guides & Exam Preparation Blueprints
Reinforce your knowledge with domain breakdowns, trap question analyses, and architectural cheat sheets:
Exam Guidelines & Simulation Rules
Frequently Asked Questions (FAQ)
The AIP-C01 exam consists of 75 advanced scenario, architecture, and coding questions over 180 minutes. Registration costs $300 USD, with a passing threshold of 750/1000 (75%).
Domains include: (1) Architecture and Design for Generative AI Solutions (~26%), (2) Model Selection, Adaptation, and Fine-Tuning (~28%), (3) Orchestration, Agents, and Tool Integration (~24%), and (4) Security, Safety, and Observability (~22%).
Bedrock Agents leverage ReAct (Reasoning and Acting) prompting to automatically decompose user prompts, generate logical reasoning traces, and execute API calls via OpenAPI schemas and AWS Lambda.
Knowledge Bases manages the end-to-end RAG pipeline, automatically chunking documents, generating vector embeddings with Amazon Titan, storing vectors in OpenSearch Serverless, and performing hybrid retrieval.
PEFT/LoRA (Low-Rank Adaptation) freezes base model weights and trains small adapter rank-decomposition matrices, reducing GPU memory requirements by 80% while preserving base capabilities.
Guardrails for Amazon Bedrock applies customizable content filters, denied topic policies, sensitive PII masking, and contextual grounding checks to detect hallucinations and block prompt attacks.
Amazon Bedrock Model Evaluation supports automatic evaluation (using metrics like ROUGE, BLEU, BERTScore) and human evaluation workflows with private workforces to grade accuracy, toxicity, and relevance.
Provisioned Throughput commits dedicated Model Units (MUs) for specific foundation models, guaranteeing processing capacity and eliminating rate-limiting (HTTP 429) during peak production traffic.
OpenSearch Serverless vector engine executes approximate nearest neighbor (ANN) searches across millions of vector embeddings with sub-second latency, providing the retrieval backbone for RAG.
Developers use LangChain and LlamaIndex SDKs to construct modular LLM application chains, manage memory buffers, build routing agents, and bind external tools to Amazon Bedrock endpoints.
The AIP-C01 certification is valid for three years from the date earned.
AIP-C01 certifies Senior Generative AI Architects, LLM Application Engineers, and Autonomous Agent Developers earning average US salaries between $160,000 and $225,000+.