+
+
+
+ ☁️ Amazon Bedrock AgentCore
+ AgentCore Platform &
Security Accelerator
+ 2-Day Hands-On Workshop — Go from architecture design to production-deployed AI agents with enterprise-grade security, identity, and observability.
+
+
+
+
+ The Challenge
+ The Problem
+ Enterprises are stuck between prototype and production. AI agent PoCs sit idle for 3–6 months while teams navigate scattered security reviews, identity integration, and infrastructure decisions.
+
+
+
⏳
+
3–6 Month Delays
+
Security reviews, compliance checks, and infrastructure decisions create months of back-and-forth before a single agent reaches production.
+
+
+
🔀
+
Scattered Reviews
+
Identity, networking, encryption, and observability are handled by different teams with no unified blueprint — leading to inconsistent implementations.
+
+
+
🚧
+
Prototype Graveyard
+
Promising agent PoCs never make it past the demo stage. Without a clear path to production, innovation stalls and momentum is lost.
+
+
+
+
+
+ Our Approach
+ The Solution
+ A structured 2-day sprint with modular, progressive deployment. Every stack builds on the last, and every decision is made with production in mind from day one.
+
+
+
🏗️
+
Modular Architecture
+
10 independent CDK stacks that compose into a complete platform. Deploy what you need, skip what you don't.
+
+
+
🔐
+
Security-First
+
KMS encryption, VPC endpoints, SCPs, and enterprise identity integration baked in from the foundation layer — not bolted on later.
+
+
+
⚡
+
2 Days to Production
+
From architecture whiteboard to deployed, observable, secure AI agents — with all the IaC, runbooks, and documentation to maintain it.
+
+
+
+
+
+ Architecture
+ Progressive Deployment Layers
+ Each layer builds on the one below. Deploy bottom-up for a complete platform, or pick the layers you need.
+
+
+
5
+
Observability
Tracing, monitoring, and alerting pipeline
+
observability
+
+
+
4
+
Runtime
Agent execution, orchestration, and A2A connectivity
+
runtime-orchestratorruntime-code-agentruntime-research-agent
+
+
+
3
+
Service
API gateway and managed memory for agents
+
gatewaymemory
+
+
+
2
+
Identity
Authentication, authorization, and enterprise SSO
+
authidentity
+
+
+
1
+
Foundation
Networking, security controls, and encryption
+
networkingsecurity
+
+
+
+
+
+ Schedule
+ Workshop Flow
+ Two focused days that take you from design to deployment.
+
+
+
Day 1 — Foundation & Design
+
Architecture decisions, identity integration, core infrastructure
+
1
Platform Architecture
Multi-account strategy, networking, security controls, KMS encryption
+
2
Identity & Auth
Cognito setup, enterprise IdP integration (Entra ID / Okta), OAuth flows
+
3
Gateway & Memory
MCP gateway with Lambda/OpenAPI targets, managed memory configuration
+
4
Design Review
Architecture review, threat modeling, deployment planning for Day 2
+
+
+
Day 2 — Build & Validate
+
Agent deployment, A2A connectivity, observability, hardening
+
5
Runtime Deployment
Deploy orchestrator agent, code agent, and research agent to AgentCore Runtime
+
6
Agent-to-Agent
Configure A2A connectivity, multi-agent orchestration patterns
+
7
Observability
OpenTelemetry tracing, CloudWatch dashboards, alerting pipeline
+
8
Security Hardening
SCPs, VPC endpoints, PrivateLink, final security review and validation
+
+
+
+
+
+ Customization
+ Customer Profiles
+ Every workshop is tailored. Select the profile that matches your starting point — we'll customize the module selection accordingly.
+
+
+
🌱
+
Greenfield
+
Starting fresh with AI agents. Need the full platform from scratch.
+
networkingsecurityauthidentitygatewaymemoryruntimeobservability
+
+
+
🔄
+
Migration
+
Moving existing agents from Bedrock Agents or custom infra to AgentCore.
+
identitygatewaymemoryruntimeobservability
+
+
+
🤖
+
Multi-Agent
+
Building complex multi-agent systems with A2A orchestration.
+
authidentitygatewayruntimeobservability
+
+
+
🏢
+
Platform Team
+
Building a shared agent platform for multiple development teams.
+
networkingsecurityauthidentitygatewaymemoryobservability
+
+
+
🛡️
+
Security-Focused
+
Hardening an existing deployment with enterprise security controls.
+
networkingsecurityauthidentityobservability
+
+
+
+
+
+ Parallel Execution
+ Team Workstreams
+ Three teams work in parallel, each owning a set of stacks. Dependencies are managed through SSM parameters published by each stack.
+
+
+
🏗️ Platform Team
+
Infrastructure & Foundation
+
+
networking — VPC, subnets, endpoints
+
security — KMS keys, SCPs, guardrails
+
auth — Cognito user pool, app clients
+
identity — AgentCore identity, IdP integration
+
observability — Tracing, dashboards, alerts
+
+
+
+
🤖 Agent Team
+
Agent Development & Deployment
+
+
gateway — MCP gateway, tool targets
+
memory — Semantic & event memory
+
runtime-orchestrator — Main orchestrator agent
+
runtime-code-agent — Code execution agent
+
runtime-research-agent — Research agent
+
+
+
+
🛡️ Security Team
+
Security Review & Hardening
+
+
Review all stack outputs for compliance
+
Validate KMS encryption configuration
+
Verify VPC endpoint policies
+
Audit IAM roles and permissions
+
Threat model the agent architecture
+
+
+
+
+
+
+ Deliverables
+ What You Take Home
+ Everything you need to operate and evolve your agent platform after the workshop.
+
+
📦
Production IaC
10 CDK stacks — fully parameterized, tested, and ready for your CI/CD pipeline.
+
🤖
Deployed Agents
Working orchestrator, code, and research agents on AgentCore Runtime with A2A connectivity.
+
🔐
Security Baseline
KMS encryption, VPC endpoints, SCPs, and enterprise identity integration — all configured.
+
📊
Observability Pipeline
OpenTelemetry tracing, CloudWatch dashboards, and alerting — ready for production traffic.
+
📋
Architecture Decision Records
Documented decisions, trade-offs, and rationale for every architectural choice made.
+
🗺️
Expansion Roadmap
Prioritized backlog for adding agents, integrations, and capabilities post-workshop.
+
+
+