Projects

Official TypeScript SDK for the DesignWear platform API.

Built an aerial soil classification system analyzing drone/satellite imagery to detect soil health conditions (healthy, dry/arid, waterlogged) using a custom 4-block CNN with Squeeze-and-Excitation attention. Trained with JAX/Flax (AdamW + cosine LR schedule, Orbax checkpointing) and deployed to edge devices via Google MediaPipe's TFLite pipeline.

A development project exploring various technologies and concepts.

Intelrepo is a Python-based project that demonstrates an AI-powered road object detection system using computer vision techniques. It utilizes technologies such as OpenCV, GTK for GUI, the COCO dataset, and R-CNN models to detect objects from images, videos, and live camera feeds. The project was developed as an experimental AI application to explore object detection and real-time image processing with CPU/GPU support.

Designwear Backend is the core API and CMS integration layer for the Designwear application. It leverages Next.js, Sanity CMS, Firebase Admin, and Razorpay for payments to deliver a scalable and robust e-commerce and content management platform.

Designwear is a modern fashion and apparel brand focused on innovative clothing design, premium quality materials, and digital-first shopping experiences. The platform showcases unique design collections, creative fashion concepts, and a seamless e-commerce experience for customers.