Shayari AI | End to End Generative AI Application Built a custom Generative AI system that creates Urdu and Hindi shayari based on user provided context. The project was developed completely from scratch, covering dataset creation, model fine tuning, backend serving, and frontend deployment. • Curated and cleaned a dataset of 10,000 plus shayaris • Fine tuned Llama 3.2 1B using Unsloth for efficient training • Built a FastAPI based inference service with CPU optimized deployment • Containerized the application using Docker and deployed on Hugging Face Spaces • Designed and deployed a lightweight frontend using HTML, CSS, and Vanilla JavaScript on Netlify Tech Stack: Python, FastAPI, Llama 3, Unsloth, Docker, HTML, CSS, JavaScript, Hugging Face, Netlify