Raghunath Rao

Title of the Talk: MLOps in the Wild: Engineering Reliable AI Across the Full Model Lifecycle

Abstract:
Building a machine learning model is the easy part. Keeping it accurate, compliant, cost-effective, and trusted in production — across years, across regulatory scrutiny, across shifting data patterns — is where most AI programs quietly fail. MLOps is the engineering discipline that bridges the gap between data science ambition and enterprise-grade AI reality, and it remains critically underinvested in organizations of every size.

This delivers an honest account of what it takes to operate AI at scale across industries including financial services and telecommunications. We cover the full production model lifecycle: automated training and deployment pipelines, real-time model health monitoring with anomaly detection and executive dashboards, drift detection strategies that reduce model degradation response time from days to hours, and data lakehouse migration patterns that modernize legacy infrastructure without disrupting downstream analytics. We examine how to build CI/CD pipelines for ML that cut release cycle times by 40% or more, how to design retraining triggers that respond intelligently to business signals rather than just data drift, and how to communicate model health to non-technical stakeholders in language they act on.

Whether standing up a first production ML pipeline or scaling an existing MLOps practice, which  will leave with a battle-tested framework and immediately applicable lessons.

Bio

Raghunath Rao Aggress is a Senior Data Scientist with 12+ years of experience building and
deploying AI and machine learning systems across financial services, telecommunications, and
enterprise technology sectors. He specializes in Agentic AI, GenAI, MLOps, and large-scale data
engineering, with deep hands-on expertise in designing autonomous AI agents, RAG pipelines, LLM fine-tuning, and end-to-end ML lifecycle management on Azure and Databricks. Raghunath is passionate about translating complex data science into high-impact business outcomes — reducing operational costs, improving organizational efficiency, and making AI systems that organizations can trust.