At Tredence Analytics, I built a LangGraph-based multi-agent system for First Abu Dhabi Bank that automates personal loan processing and KYC validation. It processes more than 15 document types, and reduced manual effort by 35–45% while improving scalability.
At Impetus Technologies, I built buyer-intent prediction models for ADP and scalable pipelines using PySpark, Databricks, and MLflow. I also designed a drift-monitoring platform for more than 100 production ML models, helping reduce unexpected model-degradation incidents by 25–30%.
At FIS Global, I built speech-to-text and speech-analytics pipelines that processed more than 50,000 calls per day. I also built an ML-based product-classification system for Google Shopping at GlobalLogic, and a RAG chatbot for GitHub Issues as a project.

