Mr. Nitin Kumar

Mr. Nitin Kumar

Designing and Governing LLM Agents: From Architecture to Ethics

Abstract:

Building autonomous AI agents that can operate reliably in the real world requires deliberate choices in architecture, deployment, coordination, risk management, and ethics. This session will explore proven architectural patterns, frameworks, and algorithms for integrating perception, reasoning, decision-making, and action, and why these outperform alternatives in production contexts. We’ll address the toughest scaling challenges in moving from prototypes to production, including reliability, cost control, and state management—alongside the unique obstacles ML teams face when experimenting at scale. Multi-agent scenarios will be unpacked, covering coordination models, communication strategies, and conflict resolution techniques that balance control with adaptability. We’ll examine how to manage the trade-offs between speed, accuracy, and safety when agents operate with incomplete or ambiguous information. The session will also dive into ethical safeguards—transparency, bias mitigation, privacy, and human oversight—that ensure agents remain aligned with organizational goals and societal values. Attendees will leave with a comprehensive, experience-based playbook for architecting, scaling, and governing LLM agents responsibly.

Profile:

 

Nitin Kumar is a data science and AI leader with over 20 years of IT experience. He currently serves as the Director of Data Science at Marriott International, where he leads the design and deployment of enterprise-scale data science and generative AI solutions across 30 global brands. A recognized expert in the field, Nitin also serves as an industry judge for prestigious awards in artificial intelligence and technology.

He holds a Master’s degree in Data Science from the University of Illinois Urbana‑Champaign. His research contributions have been published in peer-reviewed journals and presented at leading industry and academic conferences, reflecting his deep commitment to advancing the field of AI through both innovation and thought leadership.

© Copyright @ ijcaci2025. All Rights Reserved