Databricks-GenAI-Assoc Application Development Practice Question
A GenAI engineer is building a retrieval-augmented generation (RAG) application using Databricks Vector Search. They notice that the retriever sometimes returns irrelevant chunks that hurt answer quality. They want to add a reranking step to improve the relevance of retrieved documents before passing them to the LLM. Which component should they add to their RAG pipeline?
⚠ Common exam trap
The trap here is assuming that a larger embedding model or a fine-tuned LLM can fix retrieval relevance without a dedicated reranking step.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
A cross-encoder reranker
A cross-encoder reranker is specifically designed to reorder retrieved documents by jointly encoding the query and each document, producing a relevance score. This step filters out irrelevant chunks and improves the quality of context passed to the LLM. In Databricks, you can incorporate a reranker as a separate component in a LangChain or custom RAG pipeline, often using a model served via Mosaic AI Model Serving or an external endpoint.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
A larger embedding model
Why it's wrong here
Switching to a larger embedding model might improve retrieval quality slightly but does not directly rerank or filter the retrieved chunks. It also increases indexing and query costs. The problem is that the current retriever returns irrelevant chunks; a reranking step is more targeted and cost-effective than replacing the embedding model, which may not solve the relevance issue.
- ✓
A cross-encoder reranker
Why this is correct
A cross-encoder reranker scores each query-document pair jointly, capturing fine-grained interactions and reordering the initial retrieval results by relevance. In Databricks, you can integrate a reranker model (e.g., from Hugging Face) as a separate component in the RAG chain. This directly addresses the issue of irrelevant chunks by refining the top-k results before generation.
- ✗
A fine-tuned LLM
Why it's wrong here
Fine-tuning the LLM on domain data can improve generation but does not address the retrieval relevance problem. The irrelevant chunks would still be passed as context, potentially leading to hallucinations or off-topic answers. The engineer needs to improve the retrieval stage, not the generation stage, so fine-tuning the LLM is not the right solution here.
- ✗
A BM25 retriever
Why it's wrong here
BM25 is a lexical retrieval method based on term frequency and inverse document frequency. It does not consider semantic similarity and may miss paraphrases. While it can complement vector search, it does not rerank results and would not resolve the issue of irrelevant chunks returned by the vector retriever. It is typically used as a first-stage retriever, not a reranker.
About these practice questions
Courseiva writes every Databricks-GenAI-Assoc question from scratch — 330 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Databricks exam blueprint
This Databricks-GenAI-Assoc practice question is part of Courseiva's free Databricks certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the Databricks-GenAI-Assoc exam.