AI-Driven Workflow Orchestration at Enterprise Scale
summary
A concrete system is described: ingestion of build logs test results pipeline metrics and PR metadata into a model-ready telemetry layer, feature engineering that creates risk signals, anomaly detection that surfaces weak signals, and a risk scoring engine that gates pipeline stages or pauses execution. The claim is closed-loop retraining reduces false positives and prevents risky deployments, producing faster rollouts with lower business impact, reduced MTTR and fewer firefights in production.
AI-written summary · about
chapters (11)
- 0:00Introduction and Overview
- 0:50Agenda and Session Outline
- 1:25Challenges with Traditional CICD
- 3:38Understanding Predictive DevOps
- 5:43System Overview and Data Collection
- 8:53Ingestion and Telemetry Layer
- 11:11Feature Engineering and AI Analysis
- 12:38Model Layer and Anomaly Detection
- 14:59Risk Scoring Engine
- 18:24End-to-End Predictive DevOps Flow
- 22:21Operational Benefits and Conclusion
related talks
21mAI-driven optimization of CI/CD pipelines
26mAccelerating DevOps with AI
09mAI-Driven Platform Engineering
40mAI-Augmented DevOps with Platform Engineering
28mAI for Everything After Code
25mDe-risking Deployments in an AI World
30mArchitecting Integrity: Responsible Software Evolution in the AI Era
19mFrom Alerts to Action: AI-Powered Incident Response Systems