Study Notes

Appearance

AIP-C01: Exam Tips & Strategy ​

← Back to Overview

Final Preparation

This page contains strategies, tips, and common traps to help you succeed on exam day.

Exam Format Overview ​

Exam Details:

  • Duration: 170 minutes
  • Questions: ~85 questions
  • Format: Multiple choice, multiple select
  • Passing Score: 750/1000
  • Languages: English

Question Types:

  1. Single Answer - Choose the ONE best answer
  2. Multiple Answer - Choose ALL that apply
  3. Scenario-Based - Read a scenario, then answer questions

Time Management Strategy ​

  • Total Time: 170 minutes
  • Questions: ~85
  • Time per Question: ~2 minutes
  • Review Time: keep ~20 minutes at the end

Tips ​

  1. Read questions carefully
  2. Flag difficult questions and move on
  3. Answer easy ones first
  4. Reserve review time
  5. Do not get stuck on one question

Common Exam Traps ​

Trap 1: InvokeModel vs InvokeModelWithResponseStream ​

text
Need full response? → InvokeModel
Need streaming UX?  → InvokeModelWithResponseStream

Trap 2: OpenSearch Serverless vs Standard OpenSearch ​

  • Knowledge Bases / RAG → OpenSearch Serverless
  • Full-text search with custom config → managed OpenSearch cluster

Trap 3: Missing Keywords ​

Watch for:

  • "MOST cost-effective"
  • "LEAST effort"
  • "immediate"
  • "minimum changes"

Trap 4: Assumption-Based Answers ​

Do not add unstated requirements. Answer only what the scenario establishes.

Domain-Specific Tips ​

Domain 1 ​

  • Highest-frequency topics: FM selection, Knowledge Bases, chunking, private endpoints
  • Common trap: choosing fine-tuning when RAG is more appropriate

Domain 2 ​

  • Highest-frequency topics: Bedrock Agents, RAG architecture, API integration
  • Common trap: not knowing when to use Agents vs a simple Knowledge Base query

Domain 3 ​

  • Highest-frequency topics: Guardrails, IAM least-privilege, VPC endpoints
  • Common trap: confusing CloudTrail with CloudWatch

Domain 4 ​

  • Highest-frequency topics: PTU vs on-demand, token efficiency, batch inference
  • Common trap: choosing PTU for sporadic workloads

Domain 5 ​

  • Highest-frequency topics: groundedness, human vs automated evaluation, CloudWatch metrics
  • Common trap: forgetting that Groundedness is the most RAG-specific metric

← Back to Overview