Answer: Use Bedrock pre-trained FMs and apply prompt engineering

## Explanation **Correct Answer: B - Use Bedrock pre-trained FMs and apply prompt engineering** **Why Option B is correct:** 1. **Amazon Bedrock** provides access to pre-trained foundation models (FMs) from leading AI companies 2. **Prompt engineering** allows customization without model training or fine-tuning 3. **Minimal configuration** - Bedrock offers managed inference endpoints with simple API calls 4. **No model training required** - Uses pre-trained models that already understand summarization tasks **Why other options are incorrect:** **Option A (Amazon Comprehend):** - While Amazon Comprehend does offer text summarization, it's a specific NLP service with limited customization - Less flexible than Bedrock for prompt engineering and using different foundation models **Option C (SageMaker training jobs):** - Requires custom datasets and model training - Involves significant configuration, tuning, and training efforts - Contradicts the requirement of "don't want to tune or train models" **Option D (EC2 Auto Scaling groups):** - Requires hosting and managing infrastructure - Needs model deployment, scaling configuration, and infrastructure management - Involves significant operational overhead and configuration **Key AWS Concepts:** - **Amazon Bedrock**: Fully managed service for foundation models with minimal configuration - **Prompt Engineering**: Technique to guide pre-trained models without retraining - **Serverless AI**: Bedrock provides serverless inference, eliminating infrastructure management - **Pre-trained Models**: Ready-to-use models that don't require training or fine-tuning

Author: Ritesh Yadav