Abdul Rasheed Shaik

From Automation to Autonomy: Engineering Trustworthy Multi-Agent LLM Systems for Regulated Financial Enterprises

Abstract:

Financial institutions are moving beyond single-purpose AI automation toward autonomous, multi-agent systems that can decompose complex enterprise workflows, coordinate across specialized agents, and handle exceptions without constant human intervention. This shift promises significant gains in efficiency and scale, but it also introduces new engineering challenges: agent-to-agent handoff, autonomous decision boundaries, and accountability for actions no single human directly authorized.

This keynote presents a practitioner's framework for architecting agentic AI systems in production financial environments, where regulatory compliance, auditability, and system reliability are non-negotiable constraints rather than afterthoughts. Drawing on real-world experience building secure, distributed API ecosystems and open banking platforms, the session explores how classical distributed systems principles — fault tolerance, consistency, graceful degradation — must be extended to govern autonomous, reasoning-capable agents operating at enterprise scale.

Particular attention will be given to designing for accountability: how to build agentic systems that remain explainable, auditable, and recoverable when autonomous decisions go wrong, and how engineering teams can move from brittle single-agent automation to resilient multi-agent architectures without sacrificing the governance regulated industries require.

Attendees will leave with a practical mental model for evaluating when multi-agent architectures make sense, common failure modes to design around, and lessons learned from deploying agentic AI in a highly regulated, high-availability production environment.

Profile:

Abdul Rasheed Shaik is a Senior Engineering Leader with 16+ years of experience architecting and scaling cloud-native platforms and distributed systems in the financial technology domain. He leads multiple engineering teams focused on enterprise data integrations, regulatory compliance, and secure API ecosystems, with responsibility for architecture, delivery, and engineering standards. His expertise spans software architecture, distributed systems, cloud-native infrastructure, Agentic AI, and multi-agent LLM architectures. He is an IEEE Senior Member, a Fellow of IETE, and Founding Member & Vice-Chair of the ACM Richmond, VA, USA Chapter. He serves as a Program Committee member for multiple International conferences & journals in software engineering and applied AI.