Title:
From Prompted Tools to Governed Agents: Architecting Reliable AI Systems in the LLM Era

Abstract:
Recent advances in large language models (LLMs) have enabled AI systems that can reason, plan, and interact with external tools to perform complex tasks. These capabilities are driving the emergence of agentic AI systems capable of autonomously executing workflows across digital infrastructures. However, deploying such systems in real-world environments introduces important challenges related to reliability, governance, security, and operational safety.

This keynote explores the transition from prompt-driven tool usage toward governed agent architectures designed for production environments. The talk will discuss key architectural patterns for building reliable AI systems, including structured tool orchestration, policy-aware execution, schema validation, and observability frameworks. Drawing on insights from both research and industry practice, the session will highlight practical strategies for deploying trustworthy and scalable AI agents, and will conclude with a discussion of emerging trends in agent governance, self-evaluating AI systems, and multi-agent collaboration.

Profile:
Aswathnarayan Muthukrishnan Kirubakaran is a data engineering and AI systems professional specializing in large-scale data infrastructure, distributed machine learning pipelines, and intelligent automation systems. He has extensive industry experience designing and deploying production-grade data platforms and AI-driven analytics systems across enterprise technology platforms and emerging intelligent systems. His work focuses on building reliable and scalable AI systems that integrate machine learning, data engineering, and autonomous agent architectures. He has contributed to research in areas such as federated learning, distributed AI pipelines, and autonomous decision systems, and actively contributes to the technology community through technical publications, peer review activities, and mentoring initiatives.