Building Advanced Agentic Systems on AWS
Please note that this is not a prep course for any AWS certification!
This document contains links mentioned in the “Building Advanced Agentic Systems on AWS” course, and additional material on the same topics.
Module 0: Course Introduction
Full name: Building Advanced Agentic Systems on AWS
Product code: 300-MLADAS-10
These links are based on the course version 1.0.4.
Online and classroom course options: Building Advanced Agentic Systems on AWS on AWS Skill Builder
The official course outline: Building Advanced Agentic Systems on AWS (PDF)
The official course material: eVantage Bookshelf (VitalSource)
Lab environment: AWS Builder Labs
Recommended prerequisite course: Building Agentic AI with Amazon Bedrock AgentCore
The main tools: Strands Agents SDK, Strands Agents documentation, Amazon Bedrock AgentCore, Amazon Bedrock AgentCore Developer Guide
Additional material on AWS Skill Builder related to this course. Digital courses and microcredentials are free with a Skill Builder account; labs and SimuLearn need a Skill Builder subscription; AWS Jams are classroom events.
Skill Builder: Digital courses (free)
- Building Multi-Agent Systems with Strands Agents
- Implementing Standardized Agent Communication Protocols with Amazon Bedrock AgentCore
- Integrating Amazon Bedrock powered Agents with Model Context Protocol Servers using the Strands Agents SDK
- Extending MCP support for Amazon Bedrock AgentCore Gateway
- Designing lifecycle policies for AgentCore memory
- AgentCore Policy on Amazon Bedrock
- Securing Agentic Workflows: Identity Propagation with Amazon Bedrock AgentCore
- Debugging production agents with Amazon Bedrock AgentCore Observability
- Building Your Agentic Applications the Well-Architected Way
- AWS Security: Securing Generative AI on AWS
- AWS Agentic AI Demonstrated (microcredential)
- AWS Securing Agent Identities Demonstrated (microcredential)
- Advanced Prompt Caching and Response Optimization
- Amazon Bedrock AgentCore Getting Started
- AgentCore Evaluation on Amazon Bedrock
- Introduction to Agentic AI on AWS
- Model Context Protocol (MCP) - Simply explained
Skill Builder: Self-paced labs (subscription required)
- Lab - Integrate Agents with MCP Servers and Tools using AgentCore Gateway
- Lab - Secure AI Agent Access with Amazon Bedrock AgentCore Identity
- Build Context-Aware Agents with Amazon Bedrock AgentCore Memory
- Lab - Build Multi-Agent Collaboration using Amazon Bedrock
- Lab Tutorial - Create Advanced AI Agents with Strands SDK and Amazon Bedrock (BETA)
- Lab - Deploying Intelligent Agents with Amazon Bedrock AgentCore Runtime
- Lab - Enhance and Scale Agents with Amazon Bedrock AgentCore
- Lab - Monitor Agent Performance and Telemetry using AgentCore Observability
- Lab Tutorial - Building Your First Model Context Protocol (MCP) Server (BETA)
Skill Builder: Learning plans and exam prep
- Build and Deploy Agents with Strands
- Agentic AI Frameworks Deep Dive
- AWS Agentic AI Learning Plan
- Builder Labs: Agentic AI on AWS
- AWS Generative AI Developer Advanced Learning Plan (includes Labs)
- Exam Prep Plan: AWS Certified Generative AI Developer - Professional (AIP-C01 - English)
Skill Builder: Jams and game-based learning
- AWS Jam - Building Advanced Agentic Systems on AWS
- AWS Jam - Building Agentic AI with Amazon Bedrock AgentCore
- AWS SimuLearn: Build with AgentCore Memory and Gateway
- AWS SimuLearn: Observability and Cost Attribution for Multi-Agent Systems
- AWS SimuLearn: Generative AI Architect
Module 1: Multi-Agent Architecture and Communication Patterns
Websites
- Coordinate multiple agents - Strands Agents
- Agent Workflows: Building Multi-Agent Systems with Strands Agents SDK
- Graph Multi-Agent Pattern - Strands Agents
- Swarm Multi-Agent Pattern - Strands Agents
- Persist state across sessions - Strands Agents
- State Management - Strands Agents
- Create custom tools (accessing state in tools) - Strands Agents
- SessionState - Amazon Bedrock API Reference
- Sharing AgentCore short-term memory between Strands agents, travel planning example notebook (awslabs samples)
- Agents as tools - Strands Agents
- Agent-to-Agent (A2A) Protocol - Strands Agents
- AgentCore Memory Session Manager - Strands Agents
- Add memory to your Amazon Bedrock AgentCore agent
- Memory organization in AgentCore Memory
- AgentCore Memory best practices (encryption, memory poisoning, least privilege)
- Deploy A2A servers in AgentCore Runtime
- Using DynamoDB as a storage backend for Strands Agents
- What is A2A? - Agent2Agent Protocol
- Enterprise Implementation of A2A
- A2A and MCP: Detailed Comparison
- Security Best Practices - Model Context Protocol
- awslabs/agentcore-samples on GitHub
Blog posts
- Introducing Strands Agents 1.0: Production-Ready Multi-Agent Orchestration Made Simple
- Introducing agent-to-agent protocol support in Amazon Bedrock AgentCore Runtime
- Amazon Bedrock AgentCore Memory: Building context-aware agents
- Multi-agent social intelligence with Strands Agents and Amazon Bedrock
- Introducing strands-dynamodb-storage: Durable agent storage for the Strands Agents SDK
- Multi-Agent Systems with Strands Agents: from theory to swarm intelligence
- Building Strands Agents with a few lines of code: Agent-to-Agent (A2A) Communication
- Building Intelligent Multi-Agent Systems with AWS Strands SDK and Amazon Bedrock AgentCore
- How we built our multi-agent research system (Anthropic)
Videos
- AWS re:Invent 2025 - Multi-Agent Collaboration with Amazon Bedrock AgentCore for Advanced SecOps (NTA407)
- Breaking multi-agent silos: A2A + MCP in action with Strands Agents (AWS Show and Tell)
- Build Your First Agent Workflow with Strands (AWS Events)
- AWS re:Invent 2025 - Make agents remember with Amazon Bedrock AgentCore Memory (AIM331)
- Building Smarter AI Agents: Memory Management with AgentCore (AWS Developers)
- Building Multi-Agent Systems with Strands Agents: Swarm, Graph and Tool Patterns (Analytics Vidhya)
Module 2: Context Engineering and Performance Optimization
Websites
- Token-Oriented Object Notation (TOON)
- toon-format/toon on GitHub
- Format Tokenization Exploration playground
- Amazon Bedrock Prompt Caching
- Prompt caching for faster model inference - Amazon Bedrock
- Amazon Bedrock Pricing
- Amazon Bedrock model provider, including caching - Strands Agents
- Executable Code Actions Elicit Better LLM Agents (CodeAct paper, arXiv)
- Effective context engineering for AI agents (Anthropic)
- Writing effective tools for agents, with agents (Anthropic)
- Conversation Management - Strands Agents
- Lesson 8: Context engineering and context management - Strands Agents
- Connect your agent to MCP tools - Strands Agents
- Amazon AgentCore Tool Search plugin - Strands Agents
- Get started with AgentCore Memory
- Use long-term memory - AgentCore Memory
- Memory strategies - AgentCore Memory
- Retrieve memory records - AgentCore Memory
- Amazon Bedrock AgentCore Gateway: A secure AI gateway for agents, tools, and models
- Search for tools in your AgentCore gateway with a natural language query
Blog posts
- Context Engineering for Agents (LangChain)
- Effectively use prompt caching on Amazon Bedrock
- Optimizing cost and latency with Amazon Bedrock prompt caching for Anthropic models
- Transform your MCP architecture: Unite MCP servers through AgentCore Gateway
- Reduced cost, better isolation, and more resilience: Strands Agents evolves next-gen capabilities
- Stop wasting tokens with the wrong AI agent memory
- Introduction to context engineering
- From Prompt Engineering to Context Engineering: Building Agents That Actually Work at Scale
- Solving Tool Overload: How the AWS AgentCore Gateway Makes Enterprise Agents Smarter and Safer
- Never Forget a Thing: Building AI Agents with Hybrid Memory Using Strands Agents
Videos
- AWS re:Invent 2025 - What Anthropic Learned Building AI Agents in 2025 (AIM277)
- Context Engineering and Conversation Management (AWS Developers)
- AWS re:Invent 2025 - Scale agent tools with Amazon Bedrock AgentCore Gateway (AIM3313)
- Amazon Bedrock AgentCore Deep dive series: Memory (AWS Show and Tell)
- AgentCore Memory: Episodic Memory and Patterns for Production Agents (AWS Show and Tell)
- AWS Tutorial - Amazon Bedrock Prompt Caching (NamrataHShah)
Module 3: Security and Compliance Implementation
Websites
- Securing Agentic AI: The Agentic AI Security Scoping Matrix
- OAuth 2.0 authorization URL session binding - Amazon Bedrock AgentCore
- What are VPC endpoints? - AWS PrivateLink whitepaper
- Authenticating requests with AWS Signature Version 4 - Amazon S3 API
- Responsible AI Lens - AWS Well-Architected Framework
- Provide identity and credential management for agent applications with Amazon Bedrock AgentCore Identity
- Authenticate and authorize with Inbound Auth and Outbound Auth - AgentCore Runtime
- Build your first authenticated agent (with Amazon Cognito)
- Set up inbound authorization for your gateway
- Set up outbound authorization for your gateway
- Policy in Amazon Bedrock AgentCore: Control Agent Interactions
- AgentCore Policy core concepts
- AgentCore Policy conditions
- AgentCore Policy authoring guide
- What is Cedar? Cedar Policy Language Reference Guide
- Protecting your data using VPC and AWS PrivateLink - AgentCore
- Use interface VPC endpoints (AWS PrivateLink) for Amazon Bedrock AgentCore
- Configure Amazon Bedrock AgentCore Runtime and tools for VPC
- Security best practices for AgentCore Runtime
- Log Amazon Bedrock AgentCore Gateway API calls with CloudTrail
- GuardDuty AI Protection
- Generative AI Lens - AWS Well-Architected Framework
- OWASP Top 10 for Agentic Applications for 2026
- Agentic Security Initiative - OWASP Gen AI Security Project
Blog posts
- The Agentic AI Security Scoping Matrix: A framework for securing autonomous AI systems
- Four security principles for agentic AI systems
- Securing AI agents with Amazon Bedrock AgentCore Identity
- Introducing Amazon Bedrock AgentCore Identity: Securing agentic AI at scale
- Manage end-user OAuth consent for AI agents with Amazon Bedrock AgentCore
- Secure AI agents with Policy in Amazon Bedrock AgentCore
- Why Policy in Amazon Bedrock AgentCore chose Cedar for securing agentic workflows
- Enforce least-privilege authorization in multi-agent AI chains using Cedar
- Secure ingress connectivity to Amazon Bedrock AgentCore Gateway using interface VPC endpoints
- Security Hub adds AI workload protection
Videos
- AWS re:Invent 2025 - Securing agent access with Amazon Bedrock AgentCore Identity (SEC313)
- AWS re:Invent 2025 - Architecting scalable and secure agentic AI with Bedrock AgentCore (AIM431)
- AWS re:Invent 2025 - Red Team vs Blue Team: Securing AI Agents (DEV317)
- AWS re:Invent 2025 - Keep Your Agents Out of Trouble with Amazon Bedrock AgentCore (AIM3330)
- Secure Agent Authorization with OAuth 2.0 - Amazon Bedrock AgentCore (Amazon Web Services)
- What is Bedrock AgentCore Policy?
Module 4: Production Monitoring, Observability, and Evaluation
Websites
- Understand observability for agentic resources in AgentCore
- View observability data in CloudWatch (traces and spans)
- Evaluators - Amazon Bedrock AgentCore
- Strands evaluation quickstart - Strands Agents
- Evaluators - Strands Evals SDK
- strands-agents/evals on GitHub
- Observe your agent applications on Amazon Bedrock AgentCore Observability
- Get started with AgentCore Observability
- Add observability to your Amazon Bedrock AgentCore resources
- Amazon Bedrock AgentCore generated observability data
- View observability data for your Amazon Bedrock AgentCore agents
- Monitor AgentCore resources across accounts
- Generative AI observability - Amazon CloudWatch
- Using Amazon CloudWatch alarms
- Evaluate agent performance with Amazon Bedrock AgentCore Evaluations
- Evaluation types - AgentCore Evaluations
- Built-in evaluators - AgentCore Evaluations
- Custom evaluators - AgentCore Evaluations
- Custom code-based evaluator - AgentCore Evaluations
- What is OpenTelemetry?
- OpenTelemetry GenAI Semantic Conventions (GitHub)
- Observability - Strands Agents
- Traces - Strands Agents
- Strands Evals SDK - Strands Agents
- AgentCore evaluations dashboard - Strands Evals SDK
- Observe, Evaluate, and Optimize Your Agent (Amazon Bedrock AgentCore Samples)
Blog posts
- Build trustworthy AI agents with Amazon Bedrock AgentCore Observability
- Amazon Bedrock AgentCore Observability with Langfuse
- Monitor on-premises and multi-cloud AI agents with AgentCore Observability
- Operating agentic AI with Amazon Bedrock AgentCore and Datadog Agent Observability: Lessons from NTT DATA
- Amazon Bedrock AgentCore adds quality evaluations and policy controls for deploying trusted AI agents
- Build reliable AI agents with Amazon Bedrock AgentCore Evaluations
- Build custom code-based evaluators in Amazon Bedrock AgentCore
- Evaluating AI agents for production: A practical guide to Strands Evals
- Evaluating AI Agents: A production blueprint with Strands and AgentCore
Videos
- AWS re:Invent 2025 - Observability for Reliable Agentic AI with Strands SDK & OpenTelemetry (NTA406)
- AWS re:Invent 2025 - Observability for AI Agents and Traditional Workloads (COP335)
- AWS re:Invent 2025 - Improve agent quality in production with Bedrock AgentCore Evaluations (AIM3348)
Labs
- Lab: Building and Deploying Intelligent Financial Agents with Amazon Bedrock Strands and AgentCore, see the Strands Agents documentation and Host agent or tools with Amazon Bedrock AgentCore Runtime
Module 5: Well-Architected Agentic AI Systems
Websites
- AWS Well-Architected Framework
- Agentic AI - Generative AI Lens
- Operational excellence - Generative AI Lens
- Security - Generative AI Lens
- Reliability - Generative AI Lens
- Performance efficiency - Generative AI Lens
- Cost optimization - Generative AI Lens
- Sustainability - Generative AI Lens
- The pillars of the framework - AWS Well-Architected Framework
- AWS Well-Architected - Build secure, efficient cloud applications
- Agentic AI Lens - AWS Well-Architected Framework
- Observability and monitoring for agentic systems - Agentic AI Lens (AGENTOPS05)
- Testing, evaluation, and validation frameworks - Agentic AI Lens (AGENTOPS06)
- Agent goal alignment and manipulation prevention - Agentic AI Lens (AGENTSEC04)
- Legacy system integration - Agentic AI Lens (AGENTREL06)
- Reasoning and execution cost optimization - Agentic AI Lens (AGENTCOST01)
- Guidance for Multi-Agent Orchestration on AWS
- Guidance for Agentic AI Operational Foundations on AWS
Blog posts
- Announcing the updated AWS Well-Architected Generative AI Lens
- Architecting for AI excellence: AWS launches three Well-Architected Lenses at re:Invent 2025
- AI agents in enterprises: Best practices with Amazon Bedrock AgentCore
- Move your AI agents from proof of concept to production with Amazon Bedrock AgentCore
- How AgentFlo built AI sales agents with Amazon Bedrock AgentCore - Part 1
- Using Kiro to Implement the AWS Well-Architected Agentic AI Lens
Videos
- AWS re:Invent 2025 - Build a well-architected foundation for scaling generative AI and agentic apps
- AWS re:Invent 2025 - Agents in the enterprise: Best practices with Amazon Bedrock AgentCore (AIM3310)
Module 6: Course Wrap-Up
Websites
- AWS Skill Builder
- AWS Classroom training
- AWS Certification
- AWS Certification exam preparation
- AWS Certification exams
- AWS Recertification
- Schedule an Exam
- AWS Ramp-Up Guides
- AWS Workshops
- AWS Training and Certification contact form
Courses
- Building Agentic AI with Amazon Bedrock AgentCore
- Advanced Generative AI Development on AWS
- Developing Generative AI Applications on AWS
- Spec-Driven Development with Kiro
- AWS Skill Builder courses
- AWS Builder Labs