This talk cuts through the hype around “AI SRE” to show what it actually takes to automate reliability in modern Kubernetes environments. Drawing on six years of building autonomous remediation systems before AI was mainstream, it demonstrates how complexity in cloud native infrastructure drives constant escalations, slows incident response, and limits developer autonomy.
Attendees will see how a system of specialized AI agents can continuously observe cluster state, detect issues without relying on external alerts, and perform end-to-end root cause analysis with high accuracy in real-world production environments. The session walks through how these agents collaborate to investigate failures across logs, networking, deployments, and dependencies, validate conclusions, and deliver actionable remediations in a single interface.
Beyond reactive troubleshooting, the talk explores proactive reliability and cost optimization, including predictive issue detection, performance-aware rightsizing, and intelligent workload placement. Through live demonstrations, it illustrates how organizations can reduce MTTR, eliminate noisy escalations, and give both platform teams and developers a shared, contextual understanding of system health without requiring deep infrastructure expertise.
Zak Taylor is an Enterprise Account Executive at Komodor, specializing in enterprise technology sales across Kubernetes and cloud infrastructure. Based in Seattle, he brings a background in go-to-market strategy and B2B sales, with prior experience at AutoRABIT and Amazon, where he held multiple account executive roles.
He holds a degree in Finance from Penn State University and has a foundation in leadership, having led large student organizations during his time there. Outside of work, Zak is active in his community as a board chairman and wrestling coach, reflecting a strong blend of professional drive and personal leadership.
