GENERATIVE-AI-ENGINEER-ASSOCIATE · Question #8
Generative AI Engineer is helping a cinema extend its website's chat bot to be able to respond to questions about specific showtimes for movies currently playing at their local theater. They already…
The correct answer is A. Create a Feature Serving Endpoint from a FeatureSpec that references an online store synced. The showtime data is structured, frequently updated, and needs to be retrieved with very low latency for a real-time chatbot. A Databricks Feature Serving Endpoint backed by an online store (synced from the Delta table) provides sub-millisecond point lookups by location key…
Question
Generative AI Engineer is helping a cinema extend its website’s chat bot to be able to respond to questions about specific showtimes for movies currently playing at their local theater. They already have the location of the user provided by location services to their agent, and a Delta table which is continually updated with the latest showtime information by location. They want to implement this new capability in their RAG application. Which option will do this with the least effort and in the most performant way?
Options
- ACreate a Feature Serving Endpoint from a FeatureSpec that references an online store synced
- BQuery the Delta table directly via a SQL query constructed from the user's input using a text-to-
- CSet up a task in Databricks Workflows to write the information in the Delta table periodically to an
- DWrite the Delta table contents to a text column, then embed those texts using an embedding
How the community answered
(36 responses)- A75% (27)
- B14% (5)
- C8% (3)
- D3% (1)
Explanation
The showtime data is structured, frequently updated, and needs to be retrieved with very low latency for a real-time chatbot. A Databricks Feature Serving Endpoint backed by an online store (synced from the Delta table) provides sub-millisecond point lookups by location key - exactly what a real-time agent needs. This is the least-effort, highest-performance path because Databricks handles the sync and serving infrastructure. Option B (text-to-SQL on a Delta table) adds LLM inference latency and SQL generation complexity. Option C (Workflows writing to an intermediate store) introduces periodic batch lag. Option D (embedding showtimes as text and doing vector search) is semantically wrong - showtime lookups are exact structured queries, not semantic similarity searches.
Topics
Community Discussion
No community discussion yet for this question.