Tarun Kumar Chatterjee

Legacy middleware platforms like BizTalk have long powered mission-critical enterprise workflows, but as organizations scale globally, on-premises systems struggle with cost inefficiency, limited elasticity, and rigid architectures that cannot keep pace with modern data velocity or AI-driven automation. This talk presents a first-of-its-kind enterprise initiative that migrated global BizTalk workflows to a cloud-based Confluent Kafka environment while embedding agentic AI to automate complex business decisions — a dual transformation spanning infrastructure modernization and intelligent process automation. The project began with proof-of-concepts across multiple source and sink connectors to validate compatibility with existing business requirements, followed by deployment of standardized cloud infrastructure supporting 19 mission-critical jobs. A reusable, cost-optimized architecture minimized enterprise license consumption, connector sprawl, and Kafka topic overhead, while preserving high throughput through Dead Letter Queue (DLQ) handling and retry strategies. AVRO schema adoption ensured structured, version-safe message serialization, enabling Kafka to asynchronously process millions of records at scale. What distinguishes this initiative is the integration of agentic AI directly into the operational fabric of the new architecture. Using OpenAI GPT-4 with a Model Context Protocol (MCP) server, autonomous AI agents were deployed to handle deal finalization — a traditionally manual, multi-step process. These agents communicate agent-to-agent, securely invoke APIs via Apigee, and autonomously complete tasks with minimal human oversight, moving beyond conventional rule-based automation toward systems that reason over context and coordinate multi-stage workflows. The results demonstrate clear industry relevance: 40% faster processing and $100K in annual infrastructure savings, alongside a standardized blueprint now guiding future migrations across BizTalk, Service Bus, and Event Grid systems globally. Beyond immediate cost and performance gains, the project offers a reusable migration methodology for enterprises facing similar legacy modernization challenges, reducing time-to-value and de-risking large-scale cloud transitions. This talk walks through the migration strategy, architectural trade-offs in connector and topic design, lessons from running agentic AI in production, and governance considerations for autonomous agents executing business-critical tasks. Attendees will gain practical insight into combining event-streaming infrastructure with LLM-driven agentic systems, plus a framework for evaluating where agentic AI adds genuine value versus where traditional automation suffices. Intended for architects, engineering leaders, and AI practitioners modernizing legacy integration platforms while responsibly introducing autonomous AI into core operations.

Biography:

Tarun Kumar Chatterjee is a Senior lead Software Professional and Architect with 20 years of experience designing and developing enterprise applications in the United States, currently serving as a Lead Software Engineer at Presidio, USA. He specializes in cloud technologies including Microsoft Azure and AWS (IaaS/PaaS), Azure DevOps, python and cloud-native architectures. His expertise includes Machine Learning Operations (MLOps) and AI application deployment using Azure OpenAI, Cognitive Services, Cosmos DB, Functions, and related Azure services. He has delivered scalable solutions across Retail, Travel, Hospitality, E-Commerce, and Real Estate industries. His research interests include microservices architecture, MLOps, Retrieval-Augmented Generation (RAG), Agentic AI, MCP, and containerization using Docker and Kubernetes. He has extensive experience building automated CI/CD pipelines and AI powered enterprise solutions that optimize performance and reduce operational costs. He has held key roles at organizations such as Anywhere Real Estate and Cognizant Technology Solutions U.S. Corporation, demonstrating strong technical leadership in Agile environments. He holds Microsoft Azure Solutions Architect Expert and AWS certifications. He has published/submitted more than 20 research papers in SCI/Scopus-indexed journals/international conferences. He also holds a granted German patent. Additionally, he has delivered keynote speeches and industry talks at several IEEE and Springer international conferences. He serves as a reviewer for numerous Q1/Q2 journals, including those published by Elsevier and Springer. He has made significant contributions to the global research and technology community through his extensive professional service. He has served as a reviewer for numerous IEEE and Springer international conference papers, an ACM Hackathon judge, and a project judge for the Sigma Xi International Science and Engineering Fair (Regeneron ISEF), evaluating cutting-edge research and innovation. He is a Fellow of the Institution of Electronics and Telecommunication Engineers (IETE), a Senior Member of IEEE, a Fellow of Sigma Xi, a Fellow of RAPTORS, and a Distinguished Fellow of the Soft Computing Research Society (SCRS). In addition, he has served in several leadership roles, including Advisory Board Member, Technical Program Committee (TPC) Member, Co-convenor, and Session Chair for numerous IEEE, Springer, and SCRS international conferences. Mr. Chatterjee has earned numerous prestigious international accolades in recognition of his exceptional achievements in technology and innovation. His honors include the Titan Business Award (Silver) for Information Technology – AI & Automation, the Titan Innovation Award (Gold) for Best Artificial Intelligence Technology Innovation, the Noble Award (Silver) for Outstanding Achievement in Cloud Computing and SaaS, the Nobel Business Award (Gold) for Outstanding Achievement in Artificial Intelligence, and the Globee Disruptors Award (Bronze) for Artificial Intelligence Disruptor. These awards collectively demonstrate his sustained excellence and global impact in the fields of artificial intelligence, cloud computing, and technological innovation. He has also received three IEEE/Springer Best Paper Awards and the ACM Hackathon Best Judge Award, further recognizing the quality of his research and professional contributions.