Hierarchical RAG architecture scaling to 693K chunks on consumer hardware (4GB VRAM). Features 3-address routing, hybrid vector+graph fusion, and SetFit classification.

One-click Windows installer for Z-Image Turbo AI image generation. Optimized for low-VRAM GPUs (4GB+). Features Gradio web UI, automatic setup, and GGUF model support.

"Adaptive Hybrid Quantization Framework for deploying 7B+ LLMs on low-VRAM devices (e.g., GTX 1050). Features surgical block alignment and Numba-accelerated inference.

A ComfyUI Workflow for low vram users

Lightweight 6GB VRAM Gradio web app with auto-installer for running AuraFlow locally — no cloud, no clutter.

🎥 Generate high-quality videos on budget hardware with the Wan 2.2 14B Low-VRAM Workflow for ComfyUI, optimized for smooth performance and quick results.

Taiwanese Hokkien (Taigi) speech-to-text transcriber - MediaTek Breeze-ASR-26 with faster-whisper, tuned for RTX 3050 4GB low-VRAM GPUs. Gradio UI, CLI, Docker, SRT/VTT/TXT/JSON.

One-click installer for ACE-Step 1.5

Simple FP16 image upscaler for all GPUs (low-mid end users)

Contains the notebooks and workflows configured to run inference from Wan 2.2 Animate with ComfyUI on Kaggle T4 GPUs smoothly

Lightweight Stable Diffusion engine with plugin-based pipelines, VRAM-safe execution, and full 4GB GPU support.

Audit local LLM function calling and agentic reliability. Visual tool-use benchmarking for quantized models on YOUR hardware.

A privacy-first Generative AI pipeline for prototyping 3D-style game assets on consumer hardware. Optimized for low-VRAM (4GB) GPUs using PyTorch, Diffusers, and Streamlit.

Lightweight SDXL LoRA trainer optimized for 8GB VRAM GPUs. GUI with training, auto-captioning (Ollama) and image search (SearXNG).

Technical Showcase: 22B True-MoE Engine running on 6GB VRAM (GTX 1060). Demonstrates "Surgical" NF4 quantization, dynamic expert swapping, and the custom "Grace Hopper" pipeline.

A production-ready, frugal, sovereign AI system that orchestrates India's open-source language models to achieve state-of-the-art reasoning on consumer hardware through Test-Time Compute (TTC) and Cognitive Serialization.

🚀 Run modern 7B LLMs on legacy 4GB GPUs without crashes, breaking the VRAM barrier for developers facing GPU limitations.