Project — Retrieval-Augmented Generation
A resume you can interrogate.
A client-side RAG demo built on my resume: pre-computed embeddings, similarity search to find the relevant chunks, then a generated answer. Ask it something.
Hello! I'm a simple assistant that can answer questions about Mohamed Elhag based on his resume data. How can I help you today?
This is a simplified demo that runs entirely in your browser. Want the full local-LLM version with Ollama, Langchain, and Chromadb?
See the full project on GitHub →How it works
Each chunk of my resume is stored with an embedding vector. When you ask a question, the demo embeds your query, runs cosine-similarity search to retrieve the most relevant chunks, and composes an answer from them. A production RAG system would use a real embedding model, a vector database, and a full LLM — that's what the GitHub version does.