GitHub
Jeremy Barry
Data Science, Mathematics, Finance student at William & Mary.
Hello! I'm Jeremy Barry from Newburyport, Massachusetts. I study Data Science, Finance, and Mathematics at William & Mary.
I'm interested in machine learning, AI, and entrepreneurship. If any of that intrigues you, feel free to hit me up.
Technologies I'm Familiar With:
GitHub
SQL
Excel
Python
PyTorch
Pandas
NumPy
R
Matplotlib
Seaborn
Power BI
Tableau
Postgres
Supabase
Netlify
Vercel
Railway
AWS
Google Cloud
Azure
Codex
Claude
React
Expo Go
TypeScript
CapCut
Projects
Zumah
June 2026 - Present
Zumah is an ML-powered routing platform that makes AI cheaper for enterprises by matching each request to the right model for the job. We added things like KV caching and locked prefixes to maximize savings for our clients.
Engram AI
May 2026
Engram AI is an AI-powered app that optimizes in-person networking by turning conversations into organized notes, contact cards, profile matches, and follow-up reminders.
FiftyPlates
January 2026
FiftyPlates is an iOS road trip game I made in one weekend while watching the NFL playoffs. You select states when you see their license plates on a road trip, then track the drive as the map fills in.
Crawllr
August 2025 - March 2026
My very first venture. Crawllr is a social media app that allows people to track and share their adventures (think Strava + Instagram mainly for college students). We grew this to over 400 users before moving onto bigger and better things.
Marine Learning
October 2025
Built a computer vision pipeline at William & Mary’s first annual hackathon to identify benthic species from seafloor footage, using YOLO for detection and DenseNet-121 for classification. Added custom random deletion augmentation and reached about 94% test accuracy, with training accuracy around 94–95%.
Advanced Analytics for Mario Kart
September 2025
This is my first ever full-stack web app I built. I made it for fun and for my friends, using React and Firebase to track Mario Kart race results, player stats, and performance trends over time with a custom ELO rating system. You'll think it's one of the coolest things I've made if you nerd out over Mario Kart.
LiamBot - AI Assistant
December 2024
Built and deployed an AI assistant designed to respond like my little brother, originally just as a fun side project. It ended up being a great way to learn APIs, AI workflows, and full-stack development, and gave me early hands-on experience building and integrating LLM applications.
Radiation Data Analysis
August 2023
At the C-10 Research & Education Foundation, analyzed 10,000+ radiation monitoring records across 15 datasets using advanced Excel modeling. Identified a previously unknown 314% reporting variance in data reviewed by the New Hampshire Department of Health.
Customer Segmentation Analysis
March 2024
Used Python and RFM modeling to segment customer behavior for a small local business, helping group customers for retention and targeting analysis. Looking back, the visualizations were pretty rough, but it’s cool seeing that I was building clustering models as a high schooler.
Loan Payment Calculator
September 2024
Took a school project a step further by building an interactive loan repayment calculator using Pandas, Matplotlib, and Seaborn to visualize payoff timelines and long-term payment trends.