RAG Builds Real-Time Coffee Recommendation Bot
An independent developer builds a real-time coffee recommendation chatbot using retrieval-augmented generation (RAG), scraping JavaScript-heavy roaster sites with Zyte API and Scrapy, storing content in ChromaDB, and querying via LangChain with OpenAI (GPT-4 for generation, text-embedding-3-large for 3,072-dimension embeddings). The pipeline retrieves up to k=50 semantically similar items per query, enabling fresh, inventory-aware recommendations from Dak Coffee Roasters' catalog.
Key Points
- 1Uses Zyte and Scrapy to scrape JavaScript-heavy roaster sites into structured JSON for RAG.
- 2Leverages OpenAI embeddings and ChromaDB so LLMs access fresh, semantically relevant product data.
- 3Enables inventory-aware, up-to-date recommendations and full-catalog retrieval useful for product recommendation systems.
Scoring Rationale
Practical, actionable RAG tutorial with concrete scraping and embedding details, but limited novelty and single-source coverage.
Sources
Public references used for this report.
Practice with real FinTech & Trading data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all FinTech & Trading problems
