Home About Projects + Experiences Other Interests
LinkedIn GitHub

Projects + Experiences

Here are some work I've done.

Overview

The Word Hunt Solver is a full stack web application that solves the game pigeon game "Word Hunt" by finding all valid words, computing optimal scoring words, and tracking board performances relative to previous boards.

Users submit a board, receive a complete solution with scores, and view how their results compare against prior runs through percentile statistics and an interactive leaderboard.

The backend is built in Flask and exposes endpoints for board analysis, scoring, and history retrieval. A frontend presents solutions and leaderboard data in real time. Game results are stored in PostgreSQL, powering percentile calculations and record dashboards. The entire application is containerized with Docker and deployed on Fly.io, with AWS RDS used for managed database hosting.

Word Hunt Solver

Skills & Concepts Learned

  • Full stack development with Flask
  • Database management with PostgreSQL and AWS RDS
  • Containerization with Docker and Fly.io
  • Data Stuctures and Algorithms, including implementing a Trie for efficient word searching
  • Developer tooling via DataGrip for schema design and query iteration

Live App

Github Repo

Overview

Conducted research focused on optimizing statistical models for consumer behavior analysis through conjoint studies. Applied statistical techniques to improve model efficiency while maintaining accuracy, and developed tools for analyzing consumer preferences and pricing sensitivity.

Key Accomplishments

  • Reduced model runtime and storage requirements by 20x through applying statistical tests to evaluate existing and newly developed models while preserving significance levels
  • Produced accurate consumer-behavior simulations by implementing scikit-learn models to estimate part-worth utilities for large conjoint datasets
  • Enabled detailed analysis of consumer tradeoffs by developing tools that compute willingness-to-pay estimates and price sensitivities directly from part-worth utilities

Skills & Concepts Learned

  • Applied bootstrapping techniques for model evaluation
  • Performed data analysis and transformations using scikit-learn and NumPy
  • Conducted statistical analysis in Python using SciPy (scipy.stats)
  • Rapidly onboarded into an existing mid-project codebase and contributed to collaborative development

Work in Progress

Oops! This project is still a work in progress, and the repo is not yet public. Ask me if you're interested in learning more!

Overview

Worked as a software developer in the PennVet Equine Pharmacology Laboratory, building and optimizing data-processing pipelines to support large-scale chemical and pharmacological analysis. Focused on improving performance, reliability, and automation, enabling faster and more consistent research analysis.

Key Accomplishments

  • Improved script performance by 500% by identifying computational bottlenecks and refactoring inefficient data-processing logic
  • Accelerated compound profiling by automating PubChem data extraction using Selenium and Playwright, retrieving 90+ pharmacological and biochemical properties per compound

Skills & Concepts Learned

  • Web automation and scraping with Selenium and Playwright
  • Large-scale dataset handling and workflow standardization for research environments
  • Writing maintainable, reusable utilities for collaborative scientific codebases

Work in Progress

Oops! This project is still a work in progress, and the repo is not yet public. Ask me if you're interested in learning more!

Penguin Coding

I'm a big fan of penguins.