Mandar Khoje

Bio: Mandar Khoje is a Senior Engineering Manager with more than 16 years of experience building large-scale data platforms and enterprise AI infrastructure. He has held senior engineering leadership roles at Moveworks (acquired by ServiceNow), where his work centered on the data architecture underpinning enterprise agentic AI, and at AppDynamics (acquired by Cisco), where he built observability and application-performance data platforms operating at [verifiable metric, e.g., petabyte] production scale. His work focuses on the intersection of data platforms, governance, reliability, and the secure deployment of generative AI in the enterprise. He is a published author on data architecture and AI platform engineering and serves as a peer reviewer for IEEE journals and conferences

Talk Focus: Generative AI is reshaping what enterprises demand from their data platforms and quietly expanding what can go wrong. Retrieval pipelines move sensitive data at inference time, agents act with the permissions of the systems beneath them, and the scale that makes these platforms useful also makes their failures consequential. Drawing on 16+ years building data platforms for enterprise AI and large-scale observability systems, this talk surveys the security and scale concerns that emerge when generative AI meets production data infrastructure: permission-aware data access, tenant isolation, governance and auditability of AI-driven data movement, and the reliability engineering required to operate these systems at enterprise scale. The goal is a practitioner's map of the terrain where the real risks concentrate, what architectural patterns hold up in deployment, and where current practice still falls short.