I'm excited to share my latest Kaggle notebook showcasing Culinary Chameleon, a GenAI-driven toolkit that takes any recipe and transforms it into a completely new culinary style — all in pure JavaScript and Python under the hood.

What I Demonstrated:
1. Structured Output / JSON Mode: Configured the LLM prompt to return strict JSON, ensuring every response matched my schema for ingredients, instructions, and nutrition estimates.
2. Embeddings & Vector Search: Used custom chunking logic to convert recipe text into vector embeddings stored in a FAISS index for fast nearest-neighbour lookups.
3. Retrieval-Augmented Generation (RAG): Built a RAG pipeline—fetch relevant recipes, embed & index locally, then feed top results back to the LLM for style transformation.