Vedant Kolhapure.
I'm a developer who genuinely enjoys the craft; not just writing code, but thinking about why something should be built in a certain way and making sure it works reliably when real people use it. I work at the intersection of AI and backend engineering, and I find that combination really works for me: intelligent systems are only as good as the infrastructure they run on.
AI is where my curiosity runs the deepest, understanding how language models work, making their use in real products, and building systems that go beyond demos. At the same time, I take backend development seriously as a discipline as it carries the weight of the entire system: clean API design, thoughtful data flow, and code that other people can actually work with.
Outside the tech world, I'm a big reader and someone who takes fitness pretty seriously. Hitting the gym and staying fit is less of a routine and more of a reset button for me. I think staying sharp physically and mentally makes for a better engineer, and honestly, a better human.
An AI study assistant built for engineering exam prep. Ingests textbooks and lecture notes into a RAG pipeline, and answers student questions with context-grounded accuracy. Supports 5 core subjects, generates quizzes, and handles out-of-syllabus queries via web search fallback.
A machine learning-powered healthcare web application designed to predict the likelihood of epileptic seizures from EEG brain signal data. Built using an SVM model with advanced preprocessing techniques like StandardScaler and PCA, achieving 97% accuracy. Features a Flask-based interface where users can upload EEG CSV data and receive predictions with probability scores through a clean, responsive frontend.
An NLP-powered news aggregation platform that fetches real-time articles and generates concise, structured summaries automatically. Includes sentiment analysis to surface emotional framing and bias at a glance — so you know not just what happened, but how it's being told.
A production-grade CRM web application built during an internship for a real fitness business. Manages client profiles, tracks progress, handles authentication, and eliminates manual record-keeping. Currently live and in active use by the client.
Open to internships, research collabs, and interesting conversations about AI and data.