About

I am looking for an experienced Full Stack AI Developer to build an intelligent RAG (Retrieval-Augmented Generation) application. The core function of the app is to allow users to input a website URL, automatically crawl/scrape its content, and then allow the user to "chat" with that website to get accurate answers, summaries, and references based only on the crawled data. Tech Stack ● Frontend: Vue.js / Nuxt.js (Latest version preferred) ● AI/Orchestration: LangChain (Python or JS/TS) ● Database: Vector Database (e.g., Pinecone, Weaviate, or pgvector) ● LLM Integration: OpenAI (GPT-4o/mini) or Anthropic (Claude) Key Features to Implement 1. URL Crawler: A robust input field that accepts a URL, scrapes the text content (handling sub-pages is a plus), and cleans the data for processing. 2. Embedding Pipeline: Split the scraped text into chunks, generate embeddings, and store them efficiently in a vector database. 3. Chat Interface: A clean, intuitive chat UI (similar to ChatGPT) where users ask questions. 4. Source Citing: The AI responses must include "References" or "Sources" pointing back to the specific part of the scraped website where the answer was found. 5. Context Management: The bot should remember the context of the current conversation (conversational memory). Responsibilities ● Design and develop the frontend using Nuxt.js with a focus on a responsive and clean UI. ● Set up the backend API (using Nuxt Server Routes or a separate Python/FastAPI service) to handle scraping and LangChain logic. ● Implement the RAG pipeline: Scraping -Chunking - Embedding - Vector Search - LLM Response. ● Ensure the scraper handles dynamic content (SSR/SPA) effectively. ● Write clean, modular, and well-documented code. Requirements ● Proven experience building AI/LLM applications using LangChain. ● Strong proficiency in Vue.js and Nuxt.js. ● Experience with Vector Databases (Pinecone, Milvus, Weaviate, etc.). ● Familiarity with web scraping tools (e.g., Puppeteer, Playwright, Beautiful Soup, or FireCrawl). ● Ability to write prompt engineering instructions to prevent hallucinations. Deliverables ● Fully functional web application deployed on a cloud provider (e.g., Vercel, AWS, or Railway). ● Source code repository with setup instructions. ● A brief demo video or documentation on how to use the system. To Apply Please share: 1. Links to similar AI or RAG projects you have built. 2. Your preferred vector database and scraping tool for this specific use case. 3. A brief estimate of the timeline for an MVP.