Amazon Bedrock AgentCore
From Prototype to Production
A comprehensive hands-on workshop for deploying and operating highly capable AI agents securely at scale using Amazon Bedrock AgentCore.
⏱
Duration
4–6 hours
🎯
Level
Intermediate to Advanced
🛠
Format
Hands-on labs
☁️
Platform
AWS
What You’ll Build
🔐
Secure Identity & Access
Configure OAuth 2.0 and JWT-based identity with AgentCore Identity. Grant agents scoped permissions to AWS services without hardcoded credentials.
🧠
Persistent Memory Systems
Implement short-term conversation buffers and long-term semantic memory using AgentCore Memory. Build agents that learn and adapt across sessions.
📦
Containerized Deployment
Package and deploy agents using AgentCore Runtime. Implement auto-scaling, health monitoring, and blue/green deployments for zero-downtime updates.
🔍
Observability & Tracing
Set up distributed tracing with AgentCore Observability. Monitor agent reasoning chains, tool calls, and performance metrics in real time.
🌐
Gateway & Integration
Connect agents to external APIs and enterprise systems through AgentCore Gateway. Implement rate limiting, authentication, and request routing.
💻
Code Interpreter
Enable agents to write and execute Python code in sandboxed environments using AgentCore Code Interpreter. Handle data analysis and automation tasks dynamically.
Workshop Modules
Six hands-on labs taking you from setup to production deployment
01
Environment Setup & AgentCore Introduction
Configure your AWS environment, install the AgentCore SDK, and deploy your first agent. Understand the architecture and core concepts.
⏱ 30 min
02
Identity & Access Management
Implement OAuth 2.0 and JWT authentication. Configure IAM roles, resource-based policies, and AgentCore Identity for secure agent access.
⏱ 45 min
03
Memory & Context Management
Build persistent memory systems using AgentCore Memory. Implement semantic search, conversation history, and cross-session context retention.
⏱ 45 min
04
Gateway, APIs & Tool Integration
Connect agents to external services using AgentCore Gateway. Build tool definitions, handle API authentication, and implement error recovery patterns.
⏱ 60 min
05
Observability & Production Monitoring
Implement distributed tracing, custom metrics, and alerting with AgentCore Observability. Build dashboards for monitoring agent performance at scale.
⏱ 60 min
06
Production Deployment & Scale
Deploy agents to production using AgentCore Runtime. Configure auto-scaling policies, implement CI/CD pipelines, and run load tests to validate performance.
⏱ 90 min
Prerequisites
Technical Skills
- Python programming (intermediate)
- REST API concepts
- Basic AWS knowledge
- Command line proficiency
- Docker fundamentals
AWS Requirements
- AWS account with admin access
- Bedrock model access enabled
- IAM permissions for Lambda, ECS
- CLI configured locally
- S3 and DynamoDB access
Local Setup
- Python 3.11+
- Docker Desktop
- VS Code or PyCharm
- Git
- AWS CLI v2
Learning Outcomes
By the end of this workshop, you will be able to:
✓
Deploy production-grade AI agents on AWS infrastructure
✓
Implement secure identity and access patterns
✓
Build persistent memory and context systems
✓
Connect to external APIs through Gateway
✓
Monitor agents with distributed tracing
✓
Implement auto-scaling and CI/CD pipelines
✓
Use Code Interpreter for dynamic task execution
✓
Follow AWS production best practices
Technical Stack
AWS Services
- Amazon Bedrock
- AWS Lambda
- Amazon ECS / Fargate
- Amazon DynamoDB
- AWS IAM
- Amazon CloudWatch
AgentCore Components
- AgentCore Runtime
- AgentCore Identity
- AgentCore Memory
- AgentCore Gateway
- AgentCore Observability
- AgentCore Code Interpreter
Frameworks & Tools
- Amazon Bedrock · Strands SDK · Claude 3.7 Sonnet
