From Data to Decision-Makers: Architecting Trustworthy AI at Scale.

AI’s promise lies not in isolated models but in its ability to transform how organizations make decisions. Yet, the gap between experimental AI projects and enterprise-wide adoption remains wide often due to fragmented data systems, opaque models, and lack of trust in automation. This talk explores how to architect AI ecosystems that move seamlessly from raw data to trusted decisions. Drawing from real-world experience building resilient multi-cloud and enterprise systems, we will examine the critical layers of a trustworthy AI architecture from data lineage and feature governance to explainable model pipelines and human-in-the-loop feedback systems. The session will also discuss operationalizing these principles at scale: embedding observability, transparency, and ethical guardrails directly into the AI lifecycle. Attendees will learn how to bridge the divide between data science experimentation and production-grade decision systems, how to balance automation with accountability, and how to design AI platforms that leaders can trust, not just use. The result is a framework for AI that is reliable, explainable, and actionable across the enterprise.
 

I’m Ankur Sharma. I’m a technology leader with deep experience across data centers, hybrid multi-cloud networking, time synchronization, and AI-driven systems. Over the years, I’ve led engineering teams spanning control plane, data plane, API, and UI development—focusing on building resilient and scalable enterprise platforms. My work has included innovations in orchestration, observability, and private interconnection, and I hold several patents in networking and timing for multi-cloud environments. I’m currently exploring how AI can be applied to networking and synchronization to enable greater automation, intelligence, and adaptive security. Earlier in my career, I also worked on technologies in social messaging, media, and bioinformatics.