📘 Description

This project simulates an intelligent claim supervisor named Anna who manages a team of agents to process auto insurance claims for Allstate. Built with LangChain, LangGraph, OpenAI GPT-4o-mini, and persistent memory, the system evaluates policy status, classifies damage severity, and determines payment outcomes. It features a modular, node-based structure that mirrors real insurance workflows.

📘 Description

This final project showcases a complete product strategy simulation for a fictional consumer goods company, with a focus on data-driven decision-making, brand positioning, and market disruption. Delivered as part of the MSDS 440 Strategic Consulting course at Northwestern, the simulation involved iterative experimentation across pricing, media, and trade levers.

📘 Description

This project explores the use of EdgeDB to model and analyze a real-world graph-relational dataset — Yelp businesses and reviews from the Chicago metro area. Built for the MSDS 420-3 Databases course at Northwestern University, the notebook walks through schema design, data loading, and analytical queries over a semi-structured dataset.

📘 Description

This project explores how to build and analyze a graph-relational model using EdgeDB for real-world review data. It uses the Chicago Yelp dataset and focuses on businesses, users, and reviews to uncover trends in customer sentiment and reviewer behavior.

📘 Description

This project uses EdgeDB to explore graph-relational modeling and querying on a Yelp business and review dataset from Chicago. It focuses on modeling relationships among users, businesses, and reviews to uncover insights about customer sentiment and local engagement patterns.

📘 Description

This project builds on foundational database skills by executing advanced SQL queries and performing exploratory data analysis using real-world Yelp Chicago business and review datasets. It covers join logic, subqueries, filtering, aggregation, and basic statistical analysis directly within SQL.

📘 Description

This capstone assignment concludes the MSDS 420 Database Systems course with a comprehensive project that integrates database modeling, complex queries, and data-driven insights. Leveraging a normalized schema, the solution addresses advanced analytical tasks using SQL, visualizations, and narrative synthesis.

📘 Description

This project explores the use of EdgeDB to create a hybrid graph-relational schema for Yelp business and review data in Chicago. The module focuses on modeling complex relationships such as user-review-business interactions, and using EdgeQL to extract insights through recursive queries and filters.

📘 Description

This project explores the use of EdgeDB to create a hybrid graph-relational schema for Yelp business and review data in Chicago. The module focuses on modeling complex relationships such as user-review-business interactions, and using EdgeQL to extract insights through recursive queries and filters.

📘 Description

This project introduces a custom-built AI agent inspired by Warren Buffett’s investing principles. Named “Virtual Buffett,” the assistant integrates LangChain, LangGraph, Milvus vector search, and OpenAI’s GPT-4o-mini. Developed for the MSDS 442 course at Northwestern University, the agent retrieves information from Buffett’s shareholder letters and offers context-aware investment insights using a Buffett-style persona.