Building a DOE-Architected Copilot CLI Harness for Multi-Month ITSM Analysis When production fails, we jump into our terminals. But to prevent the next failure, we need to analyze months of historical context trapped inside enterprise ITSM platforms like ServiceNow and BMC Helix. Throwing 6 months of noisy ticket data directly into an LLM context window is a recipe for high cloud bills, hallucinated patterns, and "context anxiety". In this talk, you'll learn how to build a local, highly disciplined SRE Agent Harness using GitHub Copilot CLI integrated with a DOE (Directive, Orchestration, Execution) architecture. We will detail the exact data engineering pipelines required to extract, parse, and enrich months of ITSM records into clean, semantic timelines before feeding them to your agent.
Rod Anami is a Staff Site Reliability Engineer and author specializing in global cloud infrastructure, data-driven observability, and Agentic AI. As the Lead SRE at Kyndryl’s Center of Excellence, he orchestrates enterprise-scale modernization for a global client base while leading an SRE Guild spanning 40+ countries. Rod serves as the Global SRE Profession Leader, having architected the technical accreditation and enablement programs for thousands of engineers. Rod was the first-place winner of the 2025 Kyndryl AI Hackathon and has extensively experimented with Agentic AI, including the development of custom MCP (Model Context Protocol) servers to bridge LLMs with enterprise data. His expertise extends to Data Engineering, where he designs ETL pipelines and conducts exploratory data analysis on platforms like ServiceNow and CRM to drive proactive reliability. Rod is a recognized thought leader, the curator of the SRE Manifesto, and co-author of the book "Becoming a Rockstar SRE". He bridges the gap between high-level technical strategy and hands-on engineering excellence.
