At Capgemini, I architected a governed GenAI conversational analytics platform across 10+ countries, using Vertex AI/Gemini, BigQuery, Looker and Cloud Run to enable natural-language access to regulated datasets.
I also designed an agentic enterprise intelligence architecture with MCP-based tool integration and authorization services, and defined RAG patterns for secure business-context grounding. My work established responsible-AI and security controls for identity, data access, prompt-injection protection and auditability.
Across 8+ analytics programs, I defined enterprise data architecture and governance standards that reduced data rework by approximately 30%. I also optimized Python/Dataflow ingestion and KPI pre-aggregation pipelines, reducing business-critical dashboard latency by 40%.
Previously at Publicis Sapient, I architected GCP and AWS data platforms processing 150–200M+ records daily. I delivered Python/Apache Beam pipelines that generated US$40K+ in annual savings, and built Looker semantic layers and API integrations that standardized 30+ enterprise KPIs.

