Agentic Failure Triage System
LLM-powered root-cause clustering for CI/CD test failures
Built at Ciena to tackle the hardest part of large-scale hardware validation: making sense of thousands of test failures per build cycle. An agentic workflow ingests structured failure logs from Jenkins CI/CD runs, uses LangGraph to orchestrate multi-step LLM reasoning chains, and produces clustered root-cause summaries with confidence scores and suggested fix paths.