Round 1 – Take-Home Assignment (FastAPI + RAG Project) I was given a real-world backend task to complete in around 2 days. The problem statement involved building a FastAPI-based backend system where I had to use a Retrieval-Augmented Generation (RAG) approach to answer queries about hotel reviews. I extracted and chunked hotel review data from multiple PDF files. Embedded the chunks using sentence-transformers. Stored them in a FAISS vector DB and used similarity search to retrieve the best-matching contexts.