Projects

Sentimo is a hardware and software product that lets users send physical, printed messages in real time using small Bluetooth thermal printers. Through a mobile app, users can create notes, drawings, or photos that instantly print on a paired device, turning digital communication into something tangible. It is designed to make everyday messaging more personal, memorable, and emotionally impactful.

Sole developer of a mobile app that identifies cars and helps users shop for them from just a picture. Developed using Dart in the Flutter framework, integrated with YouTube Data and MarketCheck Listing APIs. Integrated Gemini model for image analysis, providing make, model, model year, and statistics for pictured cars. Published on both iOS and Android with over 10,000 active users as of early 2026. The app revolutionizes car shopping by allowing users to simply take a photo of any vehicle to instantly get detailed information and find similar cars for sale.

My first ever programming project - a simple yet engaging HTML, CSS, and JavaScript word game that challenges players to transform one word into another by changing one letter at a time. Each intermediate step must be a valid English word, creating a puzzle that combines vocabulary knowledge with logical thinking. This project represents the beginning of my programming journey and demonstrates fundamental web development skills while creating an entertaining and educational game experience for word puzzle enthusiasts.

An interactive web application that allows users to input their salary and city of origin to directly compare the cost of living between any two major metropolitan areas in the United States. Using official federal data from the Bureau of Economic Analysis (BEA), Housing and Urban Development (HUD), and Bureau of Labor Statistics (BLS), the app employs a custom formula to calculate spending power differences, clearly showing which location is more expensive and by exactly how much. This tool provides valuable insights for job seekers, remote workers, and anyone considering relocation by quantifying real purchasing power across different cities. Thousands of people have used it, with 2.8K active users at its peak.

A local-first spreadsheet visualizer that turns Excel, OpenDocument, CSV, and TSV comparison tables into a fluid, movable canvas. Lensheet supports guided table mapping, focused cross-item comparisons, card layouts, image attachments, local autosave, and self-contained XLSX exports without sending workbook data to a server.

An interactive research tool for exploring university political alignment across independent economic and social dimensions. The experience includes more than 100 U.S. institutions, evidence and confidence indicators, filtering, comparison tools, detailed university profiles, and a transparent methodology.

A local LLM compatibility tool that helps users estimate which open models can run on their computer. Users can enter GPU, CPU, RAM, unified memory, and quantization settings to compare local model fit, tokens-per-second estimates, memory requirements, and benchmark strengths across a curated model explorer.

Co-led the development of a real-estate Excel add-in from scratch, integrating AI and OCR to intelligently digitize complex T12, rent-roll, and unit-mix documents. Created a screenshot tool in Python that runs in the Windows icon-tray and uses webhooks to communicate with the Excel add-in. Developed frontend and backend using TypeScript and Python, created an installation wizard for EXE files, and certified them with an EV signing certificate and hardware encryption token. Integrated API endpoints for OpenAI, Anthropic, and other models into the Excel add-in while conceptualizing and implementing UI/UX using Figma.

Developed an Amazon Echo skill using AWS Lambda, Flask, NextJS, and Render Cloud Hosting that allows users to input any Quizlet set into a website, parses the set with Gemini to turn it into questions, and quizzes the user via their Amazon Echo. Won both the Interactive Media Track and Education Application Track at SpartaHack X, demonstrating innovation in voice-controlled learning technology. The project showcases seamless integration between web technologies and voice interfaces to create an engaging educational experience.

Placed 4th out of 56 teams in the Databricks Data + AI Hackathon 2024. Developed a comprehensive real estate advisor platform that leverages Large Language Models to provide personalized recommendations, market analysis, property valuations, and more from a Databricks dataset. The platform demonstrates advanced AI integration for real estate decision-making, combining data analytics with natural language processing to deliver actionable insights for property investors and homebuyers.

A password-protected alumni directory for Alpha Delta Phi Michigan. Search confirmed alumni by name, company, role, location, field, and available contact details through a focused, responsive interface.

Quantitative NFL power rankings that rate every team from 0 to 100 on offense, defense, and overall. A Python pipeline pulls nflverse play-by-play data on a schedule, opponent-adjusts EPA, success rate, and quarterback efficiency with a ridge regression, regresses turnover luck out of drive scoring, and publishes the results to a React leaderboard with sortable ratings, team breakdowns, and a full methodology.