Second-year CS Student · MIT Manipal
B.Tech CSE student specialising in Financial Technology at MIT Manipal. I build full-stack software and data-driven applications, from live web platforms to financial analysis tools that turn raw data into decisions people can act on.
Work
Full-stack applications and financial data tools built independently, plus academic work in data analysis and computer vision.
A full-stack student productivity platform built for the ISTE Summer School 2026 WebDev final showcase, combining an Expense Tracker, a Notes module backed by MongoDB, a Task Manager, and a Pomodoro timer in one dashboard, all running on a live Express and MongoDB backend.
A rule-based financial analysis tool powered by the Gemini API. It parses transaction history to surface personalised spending insights, detect overspending patterns, and flag subscription leakage, translating raw data into plain-language summaries a user can actually act on.
A Flask app with a single text box that routes each question through Gemini to figure out what kind of query it is before deciding what to do with it: Pokémon questions hit the PokeAPI for stats, types, and abilities, weather questions get the city extracted and passed to OpenWeatherMap for current conditions, and everything else is answered directly by Gemini. Deployed on Render with auto-deploy from main.
A machine learning app that predicts a football team's performance tier from the FIFA 23 Complete Player Dataset, comparing Random Forest and Decision Tree models on accuracy (Random Forest: 99.2%). Includes a comparison mode to see two teams' stats side by side and which rates as the stronger overall.
A binary classification neural network implemented using only NumPy, covering weight and bias initialisation, ReLU and Sigmoid activations, forward propagation, binary cross-entropy loss, backpropagation, and gradient descent. Benchmarked against Logistic Regression, Decision Tree, and Random Forest baselines on the Breast Cancer Wisconsin dataset using Accuracy, Precision, Recall, and F1 Score.
A live background-replacement pipeline built with OpenCV and NumPy. It processes webcam frames in real time, using colour masking to segment foreground from background. No GPU or ML model required.
A computer vision pipeline for detecting and tracking moving objects in live video, outputting annotated footage with highlighted object trails. Built for the ISTE MIT OpenCV Workshop.
Background
B.Tech CSE (Financial Technology) · Manipal, Karnataka
Class XI – XII (CBSE) · Faridabad
Class VII – X (CBSE) · Faridabad
Kindergarten – Class VI · Gamharia, Jamshedpur
Toolkit
Languages, frameworks, and tools I use to build and ship projects.
Activity
A look at how consistently I've been shipping code.
Recognition
The Real-Time Ball Tracking System was selected as the top submission among all workshop participants.
Ranked in the top 50 at the Cryptonite Capture The Flag cybersecurity competition at MIT Manipal.
Campus
Cloth donation drives, old age home visits, blood donation camp organisation.
Publicity & outreach for the ROS Workshop; part of SMGD and AI & Development subsystems.
Assisted in publicity for Aurora 2026 TechWeek, a multi-event technical initiative.