Generative AI Projects

State-of-the-art LLM fine-tuning, image generation, and creative AI agents engineered for the next generation of intelligence.

LocalGPT
RAG • LLMs • Ollama

LocalGPT

Fully Private Document Intelligence Platform. AI-powered system combining hybrid search, Late Chunking, and intelligent routing for on-premise RAG.

Python LanceDB LangChain Front-end: Next.js Back-end: FastAPI
Repo
CV Fine-tuning
Transfer Learning • Vision

Fine-Tuning CV Models

Advanced transfer learning on ResNet and Vision Transformers (ViT) backbones for specialized classification and detection tasks.

PyTorch ViT ResNet Front-end: Python Back-end: PyTorch
Repo
Diffusion • Stable Diffusion

VisionForge Diffusion

State-of-the-art image generation system utilizing Stable Diffusion pipelines for high-fidelity creative asset creation.

Diffusers PyTorch Gradio Front-end: Gradio Back-end: Python
Repo
Diffusion Visual
LLM Fine-tuning
LLM • Instruction Tuning

Fine-Tuning GPT Models

Domain-specific adaptation and instruction-tuning of Large Language Models to improve reasoning and contextual accuracy.

PEFT/LoRA OpenAI Transformers Front-end: Python Back-end: OpenAI
Repo
Creative AI • Diffusion

AI Game Character Designer

Generative AI tool for creating unique game characters with detailed attributes and visual assets using Diffusion models.

Stable Diffusion Gradio GANs Front-end: Gradio Back-end: Stable Diffusion
Repo
Game Designer
Foundation Models
Multimodal • Foundation Models

Foundation Models Explorer

Comprehensive exploration of multimodal foundation models (CLIP, BLIP, Segment Anything) for cross-domain AI applications.

CLIP SAM Zero-shot Front-end: Gradio Back-end: Python
Repo
Audio • Voice AI

Text-to-Speech Engine

High-fidelity audio generation pipelines and TTS fine-tuning for natural voice synthesis across multiple languages.

FastSpeech Librosa Transformers Front-end: FastAPI Back-end: Python
Repo
TTS Engine