Amol Diwakar Agade
AI-Powered Resilience in Banking: Detecting Outages, Tail Latency, and Security Violations in Real Time
Abstract:
AI-powered resilience is quickly becoming a must-have in banking. Customers expect services to work 24/7, regulators expect strong security, and engineering teams are still pushing frequent releases while modernizing older systems. In that environment, traditional monitoring like fixed thresholds, hand-tuned alerts, and separate dashboards doesn’t scale. It creates noise, misses early warning signals, and struggles with seasonal traffic patterns and complicated dependencies in distributed platforms. This keynote explains how AI-based anomaly detection can be applied as a real operational capability within DevOps and SRE, rather than treated as a side machine-learning project.
The mindset change is moving from “this number is high” to “this behavior is unusual for what’s happening right now.” We detect anomalies using context like workload patterns, seasonality, dependency health, and recent deployments or configuration changes—so teams can focus on what truly matters. I'll share a practical AIOps architecture built for regulated banks. It combines monitoring data with change and topology signals, then scores and correlates events in real time for smart triage. Since audits matter, explainability and traceability are built in, with safe rollouts and easy rollbacks. We will walk away with a blueprint to catch outages and security risks faster, reduce alert fatigue, and lower MTTD.
Profile:
Amol Diwakar Agade is a VP-level Platform & DevOps Enablement leader at Comerica Bank in Detroit, Michigan. Over the last decade, he has helped build and scale CI/CD platforms and site reliability engineering (SRE) capabilities, modernizing delivery and operations in regulated environments. Amol’s work sits at the intersection of AI and software delivery. He designs intelligent pipelines, optimizes tests with data-driven signals, and applies predictive techniques to detect failures before they impact customers. He also implements policy-as-code and AI guardrails so automation remains explainable, auditable, and safe for systems. Through DevOps and reliability transformations, he helps teams ship faster without sacrificing compliance, stability, or customer trust.
As a speaker, Amol shares frameworks and real-world patterns on operational resilience, release excellence through automation and observability, and AI/ML in DevOps—helping engineering leaders turn telemetry into decisioning and turn governance into a delivery accelerator. Amol holds an M.S. in Information Technology Management and an M.S. in Mechanical and Aerospace Engineering. He contributes to the global Intelligent Systems community as a peer reviewer for IEEE journals and as a judge of conference programs and evaluations. At ITAI 2026, he brings a practitioner’s perspective on how organizations can survive—and thrive—in the age of AI-driven intelligent systems.