This course contains the use of artificial intelligence.

Build the skills to design, secure, deploy, and operate enterprise generative AI systems on AWS through a complete architecture-first learning experience. This course takes you from business requirements and interaction channels to production-ready platforms powered by Amazon Bedrock, foundation models, retrieval-augmented generation, AI agents, enterprise data, security controls, observability, and automation.

You will begin by exploring how large organizations such as Netflix, United Airlines, Walmart, and Tesla could apply AWS Generative AI architecture to real business scenarios. You will then learn how to read a complete architecture from left to right, understand the prompt-and-response lifecycle, identify trust boundaries, map data movement, and apply the AWS Well-Architected Generative AI Lens.

The course covers web, mobile, Slack, Microsoft Teams, APIs, and Amazon Connect experiences. You will learn how to connect structured and unstructured enterprise data from Amazon S3, Aurora, RDS, DynamoDB, Redshift, SaaS platforms, internal APIs, SharePoint, Salesforce, ServiceNow, and on-premises systems.

You will design secure application layers using Route 53, CloudFront, AWS WAF, Shield, API Gateway, Lambda, ECS, Fargate, App Runner, and EKS. You will compare synchronous and asynchronous workflows, serverless and container-based architectures, stateless and stateful services, and resilient patterns for scaling, retries, timeouts, and long-running AI tasks.

A major focus is Amazon Bedrock, including model selection, Amazon Nova, Anthropic Claude, Meta Llama, Mistral, embeddings, multimodal models, structured outputs, tool calling, prompt routing, model customization, and cost-aware inference. You will build RAG systems with Bedrock Knowledge Bases, OpenSearch Serverless, S3 Vectors, Aurora PostgreSQL with pgvector, and Neptune Analytics for GraphRAG.

You will also design agentic AI systems using Bedrock Agents, AgentCore, Strands Agents SDK, LangChain, LangGraph, Step Functions, and Bedrock Flows. Topics include planning, tool execution, memory, human approval, permission boundaries, error recovery, loop prevention, and enterprise automation.

The data engineering modules show you how to create ingestion pipelines with AWS Glue, AppFlow, Database Migration Service, DataSync, EventBridge, SQS, Kinesis, Lambda, and Step Functions. You will process documents with Amazon Textract and Bedrock Data Automation, preserve metadata, select chunking strategies, generate embeddings, synchronize knowledge bases, and measure retrieval quality. You will also learn how to evaluate model responses, groundedness, safety, agent decisions, and task completion using automated metrics, LLM-as-a-judge methods, human reviews, regression datasets, and release quality gates.

Security and governance are integrated throughout the course. You will implement IAM, Cognito, least-privilege access, Bedrock Guardrails, prompt-injection defenses, sensitive-data protection, VPC isolation, encryption, audit logging, responsible AI reviews, and compliance evidence.

Finally, you will master monitoring, evaluation, CI/CD, infrastructure as code, caching, cost management, backup, disaster recovery, and multi-region resilience. Hands-on labs and a comprehensive capstone guide you through designing a secure, scalable, reliable, observable, and governed AWS GenAI platform ready for enterprise use.

By the end, you will translate business requirements into defensible architecture decisions and confidently communicate AWS GenAI designs clearly to engineering, security, risk, operations, and leadership teams.

This course is ideal for cloud architects, AI engineers, developers, security professionals, technical leaders, and anyone preparing to build production-grade Generative AI, RAG, and AI agent solutions on AWS.