Maitray Mukeshkumar Modi

Unifying Enterprise AI/ML: Architecture, Governance, and the Path from Data Fragmentation to Production

Abstract:

Enterprise AI/ML initiatives often operate across separately managed data lakes, warehouses, streaming systems, ML platforms, and business intelligence tools. This fragmentation creates integration overhead, duplicated data pipelines, governance blind spots, and training-serving skew. Surveys cited in the source indicate that data scientists spend 60–80% of their working time on data access, preparation, and quality reconciliation rather than model development.

This session examines unified cloud-based data analytics platforms as an architectural approach to these challenges, presenting a four-dimensional framework covering shared ACID-transactional storage, unified governance and lineage, batch-streaming convergence, and governance-native machine learning.

The session explores how shared storage and time-travel capabilities support reproducible ML workflows, while unified metadata and lineage improve data quality, model provenance, and regulatory documentation. It also examines how managed feature stores and converged batch-streaming architectures enable consistent feature computation across training and inference, addressing a major source of production degradation.

Attendees will examine deployment patterns across telecommunications, financial services, retail, and healthcare, followed by a structured comparison of Microsoft Fabric, Databricks Lakehouse, Snowflake Data Cloud, and Google BigQuery with Vertex AI across eight enterprise AI/ML dimensions. The session concludes with a three-phase, governance-first migration framework for sequencing governance foundations, priority workload migration, and platform consolidation.

Profile:

Maitray Mukeshkumar Modi is an accomplished AI/ML Senior Solution Architect and Technical Product Leader with more than 23 years of experience delivering enterprise-scale cloud, AI/ML, and data transformation initiatives across telecommunications, retail, and healthcare. He currently leads AI/ML and data platform modernization for a leading U.S. telecommunications organization, focusing on Generative AI, multi-agent systems, real-time analytics, and Lakehouse-based architectures.

He has deep expertise across AWS and Azure ecosystems including Amazon Bedrock, SageMaker, Azure OpenAI, Azure AI Foundry, Microsoft Fabric, Databricks, and Synapse Analytics — spanning cloud-native architecture, MLOps/LLMOps, event-driven systems, and enterprise governance. Earlier in his career, he architected large-scale AI/ML, analytics, and cloud modernization platforms across North America, with over a decade of specialized experience in telecommunications transformation involving real-time event processing and customer experience platforms.

Maitray holds a Master's in Computer Applications and a Bachelor of Commerce from Gujarat University, India, along with certifications including AWS Certified Machine Learning Engineer, AWS Certified Solutions Architect, AWS Certified AI Practitioner, Microsoft Azure Certified Data Architect, PMP, and SAFe® 4 Certified Product Owner/Product Manager.