KUSHAGRA GUPTA

• Productionized an AzureML SKU pricing engine for Tilray by integrating sales, inventory, and competitor feeds across 4M+ entries in the German pharma market; achieved R² 0.99, MAE 5.12, MAPE 1.8%, and added MLflow-based model registry, drift checks, and scheduled AutoML retraining for long-term stability • Developed an MCP-Server powered agentic AI system for high-volume EDI 277/877 parsing, processing 160K+ character multi-claim files (256 claims) via LangGraph, advanced chunking, and Azure AI Search

• Automated the end-to-end Material Reconciliation Process for Reliance Jio by integrating large-scale datasets from DPR, SAP, Parameter Dump, and PDIMS, processing over 3 million entries across 100K+ towers in 17 states. • Built an end-to-end material reconciliation pipeline with a Tkinter-based Python UI (ingestion from DPR/SAP/PDIMS, schema checks, dedupe, exception queue, audit logs); reduced reconciliation time from 16 hours to <5 minutes.