Design a chatbot over structured and unstructured data

Q: Design a chatbot over structured and unstructured data

This is a ML System Design interview question from Google for Machine Learning Engineer roles. View the full question and solution on PracHub.

Q: How do I approach ML System Design interview questions?

ML System Design questions require understanding of core concepts and practice. PracHub provides solutions with explanations to help you master ml system design interviews.

Question

Loading...

Design a chatbot that can answer user questions using both:

Structured data (e.g., relational tables such as orders, products, pricing, user accounts), and
Unstructured data (e.g., documents, knowledge base articles, FAQs, PDFs, internal wikis).

The chatbot should:

Provide accurate, grounded answers
Support follow-up questions (multi-turn)
Cite sources when possible
Respect permissions (users can only access authorized data)

Discuss:

High-level architecture and main components
Data ingestion and indexing strategy for both data types
Retrieval strategy (including when to query SQL vs retrieve documents)
LLM prompting / tool-calling approach
Safety, privacy, and access control
Evaluation (offline + online) and monitoring
Latency/cost considerations and scalability

Design a chatbot over structured and unstructured data

Solution

Comments (0)