Theme: Based on Generative AI and Machine Learning on AWS / DevOps practices.
Description: In the SRE world, "Mean Time to Recovery" (MTTR) rules all. But while the industry is caught up in the GenAI hype, how do we go from "chatbots" to production-ready autonomous systems?

My talk bridges the gap between research and real-time SRE practice. From my background as an award-winning researcher and practicing SRE, I will show how to go from simple LLM conversations to "DevOps Agents" using Amazon Bedrock and AWS Lambda. We will see how to architect a secure workflow where an AI agent doesn't just "talk," but actually analyzes CloudWatch logs, correlates multi-cloud anomalies, and writes Terraform remediation code.

As an AWS All Builders Welcome Grantee, I'm passionate about the practical and ethical application of these technologies. We will discuss the "Day 2" reality of GenAI: how to ensure your agents have the right guardrails to prevent hallucinations and maintain infrastructure integrity.

So, to be direct what I bring to table would be :
The Researcher’s Lens: Architecture patterns for secure, deterministic GenAI Agents.
The SRE’s Reality: Integrating Bedrock with existing DevOps pipelines without sacrificing reliability.
Safety & Governance: How to apply real-world guardrails to prevent AI-driven "infrastructure drift."
Multi-Cloud Context: How to use multi-cloud insights to improve AWS-specific incident response.
From my experience as Speaker from local AWS User Group, I would ensure the session must provide 'Value' for SREs and Platform Engineers looking to modernize their operational toolboxes with a foundation rooted in research-backed best practices.