I built Carbonly, a full-stack carbon estimation app that uses the Claude API to provide real-time estimates and user guidance. I designed its responsive React frontend, integrated public emissions datasets, and deployed it on Netlify.
On Razor Recon, I built an automated system that matches merchant settlement records against payout data to surface discrepancies. The project automates payout-to-ledger matching and is open-sourced under the MIT License.
I also architected an event management workflow system with role-based access control and automated task and order routing. I separated its frontend and backend for scalable, multi-tenant deployment.
During the NIT Jamshedpur Machine Learning Program, I built and deployed a production-style ML pipeline covering preprocessing, model benchmarking, evaluation, and Streamlit deployment. At Indian School of Coding, I completed full-stack training and built Python applications integrating backend logic, APIs, and deployment.
