As enterprises increasingly adopt artificial intelligence at scale, the challenge is no longer limited to model development but extends to building resilient, distributed platforms capable of delivering intelligence across complex business ecosystems. This keynote explores the evolution of enterprise AI architectures, including Retrieval-Augmented Generation (RAG), agentic AI systems, cloud-native data platforms, and real-time analytics frameworks.

Drawing from practical experiences in financial services and large-scale enterprise deployments, the session will discuss architectural patterns, governance considerations, scalability challenges, and emerging best practices for creating production-grade AI systems that support decision-making across distributed environments.

Modern enterprises in regulated industries face a critical challenge: they cannot govern what they cannot find. As distributed data systems scale to millions of columns across hybrid cloud environments, manual sensitive data classification becomes operationally impossible.

This keynote presents a production-validated framework for autonomous data governance deployed across over one million database columns in a regulated financial services environment. The classification layer utilizes three parallel engines, rule-based pattern matching, cloud-native machine learning, and Cortex AI to identify PII that traditional systems often miss. Furthermore, we introduce the Autonomous Control Plane (ACP), an event-driven architecture that translates these classifications into real-time access controls and dynamic masking. In production, this architecture resolved over 90% of database governance operations without human intervention. This session will examine the architectural decisions and lessons learned from deploying these systems in high-stakes environments.

This abstract investigates the transformative potential of agentic AI within distributed enterprise procurement frameworks, specifically targeting the complex requirements of large-scale public-sector Enterprise Resource Planning (ERP) systems. As procurement environments become increasingly decentralized, the integration of autonomous agents offers the promise of hyper-efficient, real-time supply chain management and automated contract lifecycle administration. However, the deployment of such systems in the public sector introduces critical challenges regarding operational transparency, regulatory adherence, and fiscal accountability.

This research outlines a design architecture for agentic systems that balances autonomous decision-making with mandatory "human-in-the-loop" governance protocols. We analyze mechanisms for ensuring auditability—ensuring that every automated procurement action is traceable, explainable, and compliant with statutory public-sector mandates. The paper further evaluates the trade-offs between rapid algorithmic execution and the need for rigorous risk mitigation in high-stakes, multi-stakeholder environments. Ultimately, we propose a scalable framework that empowers autonomous agents to optimize procurement efficiency while strengthening the integrity, fairness, and accountability essential to public-sector fiscal stewardship.

Financial institutions run some of the most sensitive distributed infrastructure on the planet, and a large portion of it still lives on IBM mainframes. Most Zero Trust literature assumes cloud-native architecture. That assumption breaks down when your identity boundary spans z/OS LPAR partitions, RACF security domains, AT-TLS policy layers, and hybrid cloud workloads simultaneously.

This keynote examines what Zero Trust security architecture looks like when applied to distributed financial infrastructure at the mainframe-to-cloud boundary. Drawing on production experience at a major U.S. financial institution, it covers three concrete problem areas: policy enforcement across heterogeneous security domains, certificate lifecycle management at enterprise scale, and the gap between Zero Trust theory and what legacy network stacks can actually enforce.

The talk moves past vendor frameworks and examines the real tradeoffs, where AT-TLS policy architecture maps cleanly onto Zero Trust principles, where RACF/ACF2 controls complement modern identity federation models, and where the friction points are that no architecture diagram accounts for. Attendees leave with a grounded understanding of how Zero Trust scales across distributed hybrid environments, not as a product category, but as an engineering discipline applied to systems that cannot tolerate downtime.

This abstract examines the complex intersection of artificial intelligence, financial risk assessment, and regulatory compliance within the U.S. insurance market. As insurers increasingly adopt machine learning models for pricing, they face the significant challenge of satisfying divergent state-level regulatory mandates regarding fairness, transparency, and non-discrimination. This research analyzes the technical requirements for specifying "fair" AI pricing models while ensuring they remain governable under the oversight of various Departments of Insurance.

We evaluate current methodologies for verifying that automated rating systems comply with state-specific statutes, particularly those governing disparate impact and data privacy. Furthermore, the abstract explores the necessity of robust algorithmic auditing frameworks that can bridge the gap between opaque computational outputs and legally defensible justification. Ultimately, the work proposes a multi-layered compliance architecture designed to harmonize innovation in predictive modeling with the rigorous consumer protection standards inherent in multi-state insurance regulation.

The Vice President (VP) of Software Engineering serves as a strategic leader responsible for overseeing the entire engineering organization, aligning technical initiatives with overarching business goals. This executive role demands a unique blend of high-level architectural vision, rigorous operational management, and the ability to foster a culture of engineering excellence across multiple teams. The VP manages organizational scaling, budget allocation, and the professional development of engineering leadership, ensuring that development cycles are efficient and scalable.

By bridging the gap between product requirements and technical execution, they champion innovation while maintaining robust system reliability and security standards. They actively cultivate high-performing, cross-functional teams, driving the adoption of best practices to accelerate software delivery and market competitiveness. Ultimately, this position is pivotal in shaping the company's long-term technical roadmap, ensuring infrastructure and software products remain resilient and future-proof.

This abstract investigates the technical and organizational imperatives for deploying agentic AI systems within high-stakes enterprise environments. While autonomous agents offer transformative potential for automation and decision support, their adoption remains constrained by concerns regarding non-deterministic behavior, security vulnerabilities, and a lack of systemic transparency. We define an architectural "reliability framework" for AI agents, centered on three core pillars: deterministic governance, verifiable reasoning, and robust fault-tolerance.

By moving beyond opaque, end-to-end neural models, this framework integrates symbolic logic with neural processing to ensure that autonomous actions remain explainable and strictly aligned with enterprise business logic. We further analyze the implementation of standardized, real-time "guardrails" that enforce policy compliance and provide immutable audit trails for every automated transaction. Ultimately, this research demonstrates how modular, audited agent ecosystems can provide the foundational technical trust required for complex enterprise applications, ensuring that autonomous intelligence enhances, rather than compromises, operational integrity and scalability.