Convolutional neural network that classifies thin blood smear slides as parasitized or uninfected. The model is a small VGG-style CNN trained on the NIH malaria cell dataset, exposed through a Flask web app that accepts an image upload and returns the prediction with the confidence score.
Regression model that predicts the first-innings total in an Indian Premier League cricket match given the batting team, bowling team, venue, current score, and overs bowled. Trained on ball-by-ball data from 2008 to 2019 and served behind a Flask form so anyone can plug in a live match state.
CLI that audits a website for SEO and front-end performance and prints a single-page report. Wraps Google Lighthouse, sitemap parsing, and a robots.txt linter so a non-expert can run one command and see which areas to fix first. Built during the MLH Fellowship pod project sprint.
Containerised version of the FLINT carbon-accounting engine, packaged so a researcher can run a simulation through a REST API or a CLI instead of building the C++ stack from source. Added Azure integration, GCP and Azure deployment docs, and refactored Docker builds, Kubernetes configs, and GitHub Actions CI for linting, scanning, and simulation workflows.