Software Engineering Leader, IEEE Senior Member, USA

Title of the Talk :
The Predictive Observability Pipeline: How AI is Shaping the Future of Self-Healing Systems

Abstract of the Talk:
For the last decade, observability has been about understanding what went wrong. We instrument our applications, collect traces, and sift through logs to find the needle in the haystack after an incident has occurred. But what if our systems could tell us what is about to go wrong? The next paradigm shift in cloud operations is the move from reactive monitoring to predictive observability, powered by Artificial Intelligence.
This talk explores the future of observability and the architecture of the intelligent, self-healing systems it enables. We will journey beyond traditional distributed tracing and explore how to build telemetry pipelines that are designed not just for human analysis, but as the primary data source for machine learning models. We will cover the practicalities of instrumenting applications, whether they are running in Docker containers on AWS or managed services on GCP, to produce the high-quality, structured data that AI models require.

Drawing on real-world experience building services like AWS X-Ray and leading research in applied AI, we will delve into how ML techniques can be used for advanced anomaly detection, causal inference, and failure prediction. This session will provide a visionary yet practical look at the future of our field—a future where observability doesn’t just help us fix problems faster but helps us build systems that prevent them entirely..

.