Elevator Pitch
Demos are easy; productionizing is a nightmare. Learn to treat Strands AI agents as engineering, not magic.
We’ll bridge the gap with AgentOps: versioned artifacts for Lambda/ECS, state via AgentCore, and pipelines your DevOps team will actually trust. Move from hype to scale.
Description
Let’s be honest, building a cool AI agent demo is easy now. With the Strands Agents SDK, you can have a model-driven agent running on your laptop in about ten minutes. But the moment you try to push that agent into a production environment, where it needs to be secure, observable, and patchable at 2 AM, everything falls apart. The problem is that we’re trying to shove “reasoning” agents, which are non-deterministic by nature, into our rock-solid, deterministic DevOps pipelines. Are you currently deploying agents manually? And crossing fingers that your integration doesn’t break? In this talk, we want to show you how to stop treating AI as magic and start treating it as manageable engineering. We’re going to look at the “Day 2” reality of AgentOps. We’ll share the architectural patterns we’re using to bridge this gap, specifically: The “Agent as an Artifact”: How to package your Strands logic so it’s versioned and immutable, whether you’re shipping to AWS Lambda for event-driven scale or Amazon ECS for long-running tasks. Solving the Memory Problem: We’ll dive into AgentCore to handle session identity and state, so you don’t have to rebuild the wheel every time you deploy. If you’re tired of the AI hype and want to see the actual “boring” engineering, the pipelines, the rollbacks, and the guardrails, that makes agents work at scale, this session is for you. You’ll walk away with a blueprint for building agents that your DevOps team will actually let you put in production.
Notes
This session is being delivered by two AWS Community Builders [Yeshwanth L M and Keerthivasan Kannan], who has strong expertise in AWS, DevOps and AgenticAI. Both of us have experience in delivering technical talks, keynotes and workshops for more than four years.
We are not just a builders; We are the Cloud Evangelists. Through our dedicated YouTube channels, “Tech With Yeshwanth”, “Techpechu” and “eVidhai” we have built a community of 50,000+ learners, proving our ability to simplify complex technical concepts (like AI Strands) for a wide audience. We bring the unique perspective of someone who understands both the rigors of enterprise DevOps and the educational needs of the community.
Yeshwanth L M: Speaker at AWS Student Community Day, Tirupathi 2025, AWS Community Day Kochi (AI/ML Edition) 2025 Solutions Engineer/Architect AWS User Group Leader Mysuru
Keerthivasan Kannan: Speaker at AWS Student Community Day(s), SHETECH Days, AWS UG Chennai, AWS UG Madurai Finalist of IBM AgneticAI Innovathon AI Practitioner | 7X AWS Certified