Things I've
Built

An AI-powered avatar generation system built specifically for Tamil content creators. Leverages custom fine-tuned diffusion models to produce culturally authentic digital avatars from a single photo. Features real-time rendering, voice cloning integration, and a creator-first dashboard.

A peer-to-peer GPU rental marketplace enabling AI researchers and indie developers to access affordable compute on demand. Real-time utilization dashboards, automated billing, and a WebSocket-powered live status engine.

An experimental video synthesis tool that lets creators generate cinematic B-roll from text prompts. Built on top of open-source video diffusion models with a polished editorial-style interface. Sub-second preview generation via progressive streaming.

An open-source React component library built for creative developers. Ships with physics-based animations, GSAP integrations, custom cursor cursor primitives, and WebGL-ready wrappers. 2k+ GitHub stars in its first month.

A production-ready voice interface toolkit for React. Includes waveform visualisers, speech-to-text adapters, and accessible prompt UIs. Designed for AI assistant products that need a polished voice-first experience.

A Three.js-powered portfolio builder that lets creatives showcase work in a navigable 3D gallery. Custom GLSL shaders, physics-based camera, and a no-code editor for uploading and arranging assets.