Each project below was independently built by me. Full project stories are in my portfolio.
Product development
Eve Compass, Evaradh's first product, is in daily use at Chintamani International School by 359 parent accounts and 28 teachers for 290 students (October 2026). Built role-based experiences for teachers, parents and operations staff, covering attendance, fees, records, notices and appointments.
Implemented teacher check-in with QR codes refreshed every 15 seconds and a 100-metre campus restriction, plus attendance-triggered parent notifications.
The integrated AI assistant helps teachers take attendance and post homework, adapts syllabus explanations to the class level, and guides students with hints after they share their own homework attempt.
Built with Flutter for Android and iOS, a TypeScript, Node.js and Express API on PostgreSQL with Prisma, a React admin dashboard, Firebase Cloud Messaging and the Gemini API.
Computer vision
Implemented and evaluated both architectures on MNIST using PyTorch, comparing their approaches to image recognition.
Generative models
Implemented a PyTorch GAN on MNIST with generator and discriminator networks, a custom training loop, sample grids and loss tracking. View the public notebook ↗
Robotics
Developed a ROS package for Cartesian movement and autonomous geometric drawing with a Panda manipulator.
Embedded systems
Designed and built an Arduino-based hand-gesture control system, combining machine learning with embedded motor control.