AI-Powered Cargo Tracking Tool

Designed and built the Outlook and Teams integration modules from scratch for an AI-driven cargo tracking system at a logistics company. The system reads incoming emails and Teams messages, extracts shipment data (tracking numbers, ETAs, status updates), and automatically surfaces relevant information to operations teams — replacing a fully manual process that consumed hours daily. Used LangChain and LangSmith for query processing, Dapr for service-to-service communication, and Kubernetes on GCP for orchestration.

Highlights

• 01.Built Outlook and Teams integration modules from scratch using Node.js and TypeScript, handling OAuth flows, attachment parsing, and message ingestion
• 02.Integrated LLM capabilities using LangChain to classify and process shipment data against the client's database
• 03.Deployed on Kubernetes with GCP for scalable, real-time processing across a 4-person engineering team
• 04.Implemented Dapr for microservices communication and state management