I'm building AI and data platforms at NielsenIQ, including the Discover Arthur chatbot platform, where I designed an LLM-based intent classification system achieving approximately 98% accuracy across 1,700+ evaluation scenarios.
I built its evaluation and quality-gate framework, replacing more than 90% of manual accuracy validation, and architected centralized LLM orchestration with routing, retries, streaming, error handling, and observability. I also lead multi-country retail data-pipeline enhancements across 10+ markets using Apache Beam, Dataflow, and BigQuery.
Across 15+ years, I've led enterprise delivery and technical direction at Wolters Kluwer, Cognizant, Mindtree, and other technology organizations. I bring hands-on experience designing Spring Boot microservices, distributed systems, data validation, cloud deployments, and reliable production platforms.

