I've spent seven years at Asian Paints, progressing from analytics into architecting production GenAI systems used across the company.
I built Agent SAI, a multi-agent sales intelligence platform that unifies eight-plus enterprise datasets with natural-language querying, RAG, analytics, reporting agents, and RBAC for 150+ users.
I also built FinSathi, a Teams and WhatsApp finance assistant with SAP workflows and NL-to-SQL, adopted by 1,000+ users and reducing manual reporting time by about 60%. My work includes multimodal invoice-fraud detection and an XGBoost recommendation engine that delivered roughly 2% material-cost reduction on targeted SKUs.
Across manufacturing, I’ve delivered anomaly detection, batch-quality forecasting, digital twins, data pipelines, and operator recommendations. I take end-to-end ownership from data engineering through FastAPI deployment, using Python, SQL, PySpark, Azure, Databricks, and modern LLM tooling.
