These are RAG questions of the kind Cognizant actually asks — the patterns that keep coming back in Cognizant's AI project interviews, where what matters is applying RAG sensibly for clients rather than research depth. Treat this page as a mock interview: say every answer out loud before revealing it. If one surprises you, the lesson behind it is linked at the bottom.
Cognizant RAG concept questions
List the steps of a RAG pipeline in order, and say which run offline versus per query.
Asked in
How do you decide chunk size and overlap for a new corpus?
Asked in
Does RAG eliminate hallucination? Explain your answer.
Asked in
Cognizant RAG applied & hands-on questions
Find the three production bugs in this RAG function.
Asked in
def answer(question, folder="./policies"):
docs = load_documents(folder) # 1
col.add(documents=[d["text"] for d in docs]) # 2
hits = col.query(query_texts=[question], n_results=3)
context = "\n".join(hits["documents"][0]) # 3
return llm(f"{context}\n\nQuestion: {question}")How to use this page: Cognizant rarely asks something you've never seen — they ask a standard RAG concept and then push one level deeper ("why?", "what would you do if..."). Master the concept in the RAG course lessons, and the follow-up stops being scary.
Keep practising: What Is RAG?, Chunking, Retrieval Techniques and Evaluating RAG cover what most Cognizant RAG rounds test — and the Cognizant questions section covers the rest of their selection process.