Peter Jausovec, Solo.io
Tired of spending hours troubleshooting certificate rotation failures, load balancer misconfigurations, and database connection issues? This session introduces AI Reliability Engineering (AIRE) - a framework of specialized AI agents designed to automate incident response and reduce SRE toil.
Learn how to build three core agents that form a modern reliability engineering backbone: a Terraform agent that generates configurations following AWS practices, a GitOps agent that manages complete PR workflows from creation to deployment, and an infrastructure validation agent that verifies post-deployment resources.
The talk covers implementation details any platform engineering team can adopt, including agent instruction design, MCP server integration, and testing strategies. You'll see a demo of how these agents work together to potentially save you hours of manual work.
Peter Jausovec is an engineer at Solo.io with over 17 years of experience spanning software development, QA, and engineering leadership. A recognized expert in cloud-native technologies, he specializes in Kubernetes, Istio, and AI infrastructure. Peter is a maintainer of the kagent project and has been at the forefront of AI gateway and agents development at Solo.
His recent work bridges traditional cloud-native architectures with emerging AI workloads, helping organizations navigate the intersection of service mesh, API management, and artificial intelligence.
author = {Peter Jausovec},
title = {From 4 Hours to 8 Minutes with {AI} Agents That Transform {SRE} Incident Response},
year = {2025},
address = {Dublin},
publisher = {USENIX Association},
month = oct
}
