AIP-C01: AWS Generative AI Developer Last-Minute Refresher

← Back to Overview

Final Review

This page is designed for the final "cram" session before stepping into the AIP-C01 exam.

Domain 1: FM Integration, Data Management, and Compliance (~31%)

FM Selection Quick Guide

FM	Vendor	Best For
Claude	Anthropic	Reasoning, large context (up to 200k tokens)
Llama	Meta	Open-source, fine-tuning
Mistral	Mistral AI	Efficient, high performance for size
Titan	AWS	Embeddings, summarization, AWS-native

RAG Pipeline

• S3 → chunking → embedding → OpenSearch Serverless
• Knowledge Bases manage this pipeline fully managed
• pgvector on Aurora PostgreSQL = alternative vector store
• Chunking options: fixed-size, semantic, hierarchical

Domain 2: Implementation and Integration (~26%)

API Quick Reference

Which API?
├─ Full synchronous response → InvokeModel
├─ Streaming / low latency → InvokeModelWithResponseStream
└─ Multi-step agentic workflow → InvokeAgent

Comparison Table

Feature	Knowledge Base (RAG)	Bedrock Agents
Use Case	Document Q&A, retrieval	Multi-step workflows, tool use
FM + Retrieval	Yes	Yes (optional)
External API calls	No	Yes
Best For	Static knowledge lookup	Dynamic reasoning + actions

Domain 3: AI Safety, Security, and Governance (~20%)

Guardrails Quick Facts

• Applies to both inputs and outputs
• Features: content filters, PII redaction, denied topics, word filters
• Must be explicitly applied via guardrailIdentifier in API calls
• Does NOT automatically apply to all Bedrock calls

Domain 4: Operational Efficiency and Optimization (~12%)

Provisioned Throughput vs On-Demand

	Provisioned Throughput	On-Demand
Traffic	Predictable, consistent	Sporadic, unpredictable
Pricing	Fixed (per MU/hour)	Pay per token
Commitment	1 or 6 months	None
Best For	Production 24/7	Dev/test, bursts

Domain 5: Testing, Validation, and Troubleshooting (~11%)

Model Evaluation Metrics

Metric	What It Measures
Groundedness	Is the response grounded in the retrieved context?
Relevance	Is the response relevant to the user's question?
Accuracy	Is the response factually correct?
Fluency	Is the response well-written and natural?

Key Service Decision Rules

• "Predictable throughput" → Provisioned Throughput
• "Content moderation / filter" → Guardrails for Amazon Bedrock
• "PII detection/redaction" → Guardrails
• "Vector storage for RAG" → OpenSearch Serverless (or Aurora pgvector)
• "Multi-step reasoning / tool use" → Bedrock Agents
• "Audit logging / compliance trail" → AWS CloudTrail
• "Operational monitoring / dashboards" → Amazon CloudWatch
• "Private FM access, no internet" → VPC Endpoint (PrivateLink)
• "Low latency / streaming response" → InvokeModelWithResponseStream

Common Exam Traps

Watch Out!

• InvokeModel = full synchronous; InvokeModelWithResponseStream = chunk-by-chunk
• Guardrails filter BOTH inputs AND outputs
• CloudTrail = audit/compliance; CloudWatch = operational metrics
• OpenSearch Serverless is the go-to for Knowledge Bases
• On-demand is better for sporadic traffic — don't default to PTU

---

← Back to Overview | Exam Tips