Hire Madhu N. - AI Engineer - New York

* Building an LLM-powered “Mapping” system to automate source-to-target field mappings across APIs and flat-file * integrations as part of New York Life’s core modernization from legacy mainframe systems to the FAST policy * administration platform.

Client: HDFC Bank, Gurgaon - Designed and executed an enterprise-scale Retrieval-Augmented Generation (RAG) platform integrating two major data sources—internal PDF repositories and thousands of public web pages—to power intelligent document search and chatbot use cases across HDFC. - Integrated Autogen and Anthropic’s Model Context Protocol (MCP) to build agentic interactions with the RAG backend, enabling advanced conversation workflows and multi-step task execution. - Architected schema and vector search infrastructure using PostgreSQL with pgvector extension for storing and querying embeddings; ensured scalable, real-time similarity search capabilities. - Developed reusable microservices in FastAPI, incorporating Pydantic for data validation, SQLAlchemy for ORM, and Alembic for migrations; exposed endpoints for document upload, job triggering, and admin actions via a UI.

Designed and implemented a Promotion Effectiveness Measurement System for retail clients, processing over 100,000 SKUs daily using sales, price, and promotion data. - Built a robust data ingestion and transformation pipeline using Snowflake to store and manage daily data feeds; orchestrated workflows with Apache Airflow for reliable scheduling and monitoring. - Developed an algorithm leveraging Apriori association rules to identify halo and cannibalization effects among product categories, enabling accurate measurement of promotion impact. - Delivered actionable insights to business users through dashboards and reports, quantifying the true financial effectiveness of promotions and enabling data-driven marketing and sales decisions. - Designed and built a scalable, multi-tenant Forecasting Engine based on LightGBM to generate accurate sales forecasts for both brick-and-mortar and e-commerce clients. - Integrated the forecasting system into the company’s SaaS platform using FastAPI microservices, enabling dynamic forecasting on demand per client context.