Projects

Lightning-fast quantization toolkit for PyTorch models. Supports INT8, INT4, and mixed-precision quantization with minimal accuracy loss.

The All-in-One Local AI Data Cleaning Library. Privacy-first, zero-cost data cleaning using local small language models (SLMs) like Phi-3, Qwen, and Gemma. No GPU or API keys required.

An AI-driven application for recoloring grayscale and faded images, enhancing visual quality with precision and a user-friendly interface.

A Python-based data augmentation library for specialized domains like medical imaging, improving model robustness and performance in niche applications.

An AI-powered assistant that provides tailored recommendations for furniture, color schemes, and room arrangements based on preferences and layout.