Agentic Build Day and Community Activity

Summary Report: “Agentic Build Day and Community Activity”

Event Objectives

• Provide a 3-hour technical deep-dive into AWS’s agentic services
• Walk through a real-world reference architecture: an AI-powered marketing campaign platform that orchestrates multiple specialized agents to automate audience segmentation, campaign creation, and promotions end to end
• Give attendees a working mental model of how AWS agentic services fit together
• Combine technical content with a hands-on community activity (making no-sew plush toys for children)

Speakers & Hosts

• Diem My / Dinh Nguyen – Community Activity hosts
• My Nguyen – Why Agentic AI & the MarTech use case
• Dai Truong / Thi Nguyen – Kiro Workshop Guidance
• AWS GenAI Builders Club team – technical sessions

Agenda Timeline

Key Highlights

Community Activity – No-Sew Plush Toys for Children

• A warm, hands-on opener where attendees crafted no-sew plush toys to donate to children
• A reminder that a builders’ community is about people, not just technology

My Nguyen – Why Agentic AI & the MarTech use case

• Why agentic AI is different from traditional prompt-response models
• Introduction to the MarTech reference use case: an AI-powered marketing campaign platform
• How multiple specialized agents collaborate to automate audience segmentation, campaign creation, and promotions

AWS Agentic Services in Action

• AgentCore Runtime: the execution environment for running agents
• Memory: how agents retain and recall context across interactions
• MCP Gateway: connecting agents to tools and data through the Model Context Protocol
• Agent-to-Agent (A2A) Communication: how specialized agents coordinate with each other

The Strands Agents Framework

• Overview of the Strands Agents framework
• How it ties the agentic services together into a cohesive workflow
• The role of orchestration in multi-agent systems

Full Workflow Walkthrough

• Tracing the code end to end — from a user prompt to a created campaign
• Seeing how the reference architecture executes in practice

Security Patterns, Session Persistence & Infrastructure as Code

• Security patterns for agentic workloads (IAM, least privilege)
• Session persistence to maintain state across agent runs
• Infrastructure as Code (CDK) to make the architecture reproducible

Key Takeaways & Source Code Tour

• A recap of the core concepts and how the pieces connect
• A guided tour of the full source code for self-paced exploration

Dai Truong / Thi Nguyen – Kiro Workshop Guidance

• Guidance on using Kiro for the workshop and building agentic features
• How to apply the spec-driven workflow to real projects

Key Takeaways

• Agentic AI moves beyond single prompts toward systems of collaborating agents
• AWS provides building blocks — AgentCore Runtime, Memory, MCP Gateway, A2A — that fit together
• The Strands Agents framework is the glue that orchestrates multi-agent workflows
• Production agentic systems still need security, session persistence, and IaC to be reliable
• Having a complete reference architecture + source code is the fastest way to learn by adapting

Applying to Work

• Use the MarTech reference architecture as a template for any multi-step automation use case
• Experiment with AgentCore Memory to give internal assistants persistent context
• Connect existing tools through an MCP Gateway instead of hardcoding integrations
• Apply A2A communication to break a complex single-agent prompt into specialized agents
• Adopt CDK / Infrastructure as Code so agentic deployments are reproducible and reviewable
• Try Kiro for spec-driven development of agentic features

Prerequisites

• Familiarity with AWS (Lambda, IAM, CDK) and basic Python
• No prior experience with AI agents required

Event Experience

Attending Agentic Build Day and Community Activity on May 29, 2026 was a great mix of meaningful community work and a focused technical deep-dive. Key moments included:

Starting with the community

• The event opened with a no-sew plush toy activity for children, hosted by Diem My and Dinh Nguyen — a heartwarming way to connect with other builders before diving into code.

Learning the agentic stack end to end

• My Nguyen framed the day with a clear MarTech use case, making the abstract idea of “agentic AI” concrete.
• The deep-dive into AgentCore Runtime, Memory, MCP Gateway, and A2A finally connected services I had only read about separately.
• Seeing the Strands Agents framework tie everything together clarified how orchestration actually works in practice.

Following real code

• The full workflow walkthrough — from user prompt to campaign creation — was the most valuable part, turning theory into a traceable execution path.
• The session on security, session persistence, and IaC was a good reminder that demos are easy but production needs guardrails.

Hands-on with Kiro

• Dai Truong and Thi Nguyen wrapped up with Kiro Workshop Guidance, showing how to apply a spec-driven workflow to build agentic features.

What I left with

• A working mental model of how AWS agentic services fit together
• A reference architecture I can adapt for my own use cases
• Access to the full source code to explore at my own pace

Some event photos

Overall, Agentic Build Day balanced heart and engineering — giving back to the community in the first hour, then handing us a complete, adaptable blueprint for building real agentic systems on AWS.