At Motivity Labs Inc (Client: Google), I architect and operate a Generative and Agentic AI platform on Google's Gemini Enterprise Agent Platform, serving 8,000+ users. I productionized a multi-agent architecture using CrewAI across 100K+ enterprise documents.
I also engineered NL-to-SQL and RAG pipelines, and built automated LLM quality evaluation using user feedback and false-positive and false-negative analysis. Apache Airflow pipelines handle ingestion, transformation, vector indexing, and daily corpus refreshes.
At Johnson Controls, I designed a Python ETL pipeline that transformed unstructured KPI metrics into Snowflake tables and Power BI dashboards. I also automated ingestion and preprocessing workflows and built an invoice document-processing solution using Azure AI services and OCR.
At Stacts Pvt Ltd, I built an intelligent document-processing pipeline for clinical records, including named entity and relation extraction. Automating extraction and classification workflows reduced manual labor by 65%; I deployed the NLP pipeline on Azure Cloud.
