I build the data engines that empower veterinarians and livestock owners to make faster, smarter health decisions. By transforming noisy, high-volume sensor streams into "digital biomarkers," I help bridge the gap between raw data and life-saving diagnostics. My latest project features a robust cleaning and engineering pipeline for the AcTBeCalf dataset, turning 2.4 million rows of movement data into actionable alerts for animal welfare and illness detection.

This project is a recipe recommendation system that is designed to solve a very common real-world problem. It uses a combination of data preprocessing, natural language processing (NLP), and vector similarity to suggest recipes based on a user's entered ingredients. The core of the system is a Word2Vec model trained on a dataset of diverse recipes, which is used to find semantic similarities between ingredients.

Click here to experience this project.