Discuss Databricks Databricks Certified Generative AI Engineer Associate Exam Topic 3 Question 2

A Generative AI Engineer is creating an LLM-powered application that will need access to up-to-date news articles and stock prices.

The design requires the use of stock prices which are stored in Delta tables and finding the latest relevant news articles by searching the internet.

How should the Generative AI Engineer architect their LLM system?

AUse an LLM to summarize the latest news articles and lookup stock tickers from the summaries to find stock prices.

BQuery the Delta table for volatile stock prices and use an LLM to generate a search query to investigate potential causes of the stock volatility.

CDownload and store news articles and stock price information in a vector store. Use a RAG architecture to retrieve and generate at runtime.

DCreate an agent with tools for SQL querying of Delta tables and web searching, provide retrieved values to an LLM for generation of response.

Suggested Answer: D

To build an LLM-powered system that accesses up-to-date news articles and stock prices, the best approach is to create an agent that has access to specific tools (option D).

Agent with SQL and Web Search Capabilities: By using an agent-based architecture, the LLM can interact with external tools. The agent can query Delta tables (for up-to-date stock prices) via SQL and perform web searches to retrieve the latest news articles. This modular approach ensures the system can access both structured (stock prices) and unstructured (news) data sources dynamically.

Why This Approach Works:

SQL Queries for Stock Prices: Delta tables store stock prices, which the agent can query directly for the latest data.

Web Search for News: For news articles, the agent can generate search queries and retrieve the most relevant and recent articles, then pass them to the LLM for processing.

Why Other Options Are Less Suitable:

A (Summarizing News for Stock Prices): This convoluted approach would not ensure accuracy when retrieving stock prices, which are already structured and stored in Delta tables.

B (Stock Price Volatility Queries): While this could retrieve relevant information, it doesn't address how to obtain the most up-to-date news articles.

C (Vector Store): Storing news articles and stock prices in a vector store might not capture the real-time nature of stock data and news updates, as it relies on pre-existing data rather than dynamic querying.

Thus, using an agent with access to both SQL for querying stock prices and web search for retrieving news articles is the best approach for ensuring up-to-date and accurate responses.

Get Premium Databricks Certified Generative AI Engineer Associate Questions as Interactive Web-Based Practice Test or PDF Real Exam Format Detailed Explanations Instant Access Get Your Practice Test

35 comments · newest first

Franklyn 10 months ago

Wait, can an LLM really handle real-time updates effectively?

Salley 11 months ago

A is too simplistic, it won't capture the full picture.

Della 11 months ago

C is interesting, but storing all that data sounds heavy.

Elly 11 months ago

I think B could lead to better insights on stock volatility.

Hershel 11 months ago

Option D seems the most practical for real-time data access.

Carman 12 months ago

I vaguely recall a similar question where we discussed summarizing articles. Option A seems like it could work, but I wonder if it’s the most efficient method.

Janae 12 months ago

I feel like querying the Delta table directly, like in option B, could be effective, but I’m not confident about how the LLM would fit into that.

Richelle 12 months ago

I'm not entirely sure, but I remember something about using RAG architectures. Option C might be the right approach for combining retrieval and generation.

Craig 1 year ago

I think option D sounds familiar; it seems like it aligns with what we practiced about integrating different tools for data retrieval.

Pamella 1 year ago

Option B seems like it could be a good way to go, especially if the stock price volatility is a key factor in the application. Using the LLM to generate a search query to investigate the causes of the volatility could provide some valuable insights. I'd want to make sure the LLM is well-trained on financial data, though.

Carole 1 year ago

I think option C is an interesting approach. Building a vector store to hold the news articles and stock data, then using a RAG model to retrieve and generate the response, could provide a nice balance of speed and flexibility. The key would be ensuring the vector store is kept up-to-date and the retrieval process is optimized.

Oren 1 year ago

Hmm, I'm a bit unsure about this one. The options all seem reasonable, but I'm not sure which one would be the most efficient or effective approach. I might need to do some more research on the different architectural patterns mentioned to decide which one would work best for this use case.

Lonny 1 year ago

This looks like a classic information retrieval and integration problem. I'd go with option D - using an agent with SQL and web search tools to retrieve the necessary data, then feeding that into an LLM for generation. That way, we can leverage the strengths of both the structured data access and the language modeling capabilities.

Della 2 years ago

I'm torn between options C and D. Both sound like they could work well, but I'm curious to see how the performance would compare.

Misty 2 years ago

↳ replying to Della

Agreed, performance testing would definitely help in making the final decision.

Dana 2 years ago

↳ replying to Della

It would be interesting to test both options and see which one performs better.

Cathrine 2 years ago

↳ replying to Della

Option D might provide more flexibility in querying and searching for data.

Vivan 2 years ago

↳ replying to Della

Option C could be more efficient in terms of retrieval and generation.

Christiane 2 years ago

Haha, I bet the Generative AI Engineer is hoping the LLM doesn't start buying and selling stocks on its own! Option D seems the way to go.

Alton 2 years ago

Option B is interesting, using the LLM to investigate the causes of stock volatility. That could provide some valuable insights.

Patria 2 years ago

↳ replying to Alton

C: It's definitely a unique way to leverage LLM technology for financial analysis.

Veta 2 years ago

↳ replying to Alton

B: I agree, it could help identify the reasons behind the fluctuations.

Rosio 2 years ago

↳ replying to Alton

A: Option B sounds like a smart approach to analyze stock volatility.

Ciara 2 years ago

I like the idea of a RAG architecture in option C. Storing the data in a vector store and retrieving it on the fly could be really efficient.

Selma 2 years ago

↳ replying to Ciara

Definitely, having a well-structured architecture is key for smooth operation.

Elliott 2 years ago

↳ replying to Ciara

It's important to have a system that can quickly access the necessary data.

Graciela 2 years ago

↳ replying to Ciara

I agree, using a RAG architecture could streamline the process.

Mona 2 years ago

↳ replying to Ciara

Option C sounds like a solid choice for efficient data retrieval.

Lavelle 2 years ago

I agree with Krissy. Option A seems like the most practical approach for the Generative AI Engineer to architect their LLM system.

Laurel 2 years ago

Option D seems like the most comprehensive approach. Using an agent with specific tools to handle the data retrieval and then feeding that to the LLM for generation sounds like a solid architecture.

Mozell 2 years ago

↳ replying to Laurel

C) Download and store news articles and stock price information in a vector store. Use a RAG architecture to retrieve and generate at runtime.

Reuben 2 years ago

↳ replying to Laurel

D) Create an agent with tools for SQL querying of Delta tables and web searching, provide retrieved values to an LLM for generation of response.

Lizbeth 2 years ago

↳ replying to Laurel

A) Use an LLM to summarize the latest news articles and lookup stock tickers from the summaries to find stock prices.

Latanya 2 years ago

↳ replying to Laurel

D) Create an agent with tools for SQL querying of Delta tables and web searching, provide retrieved values to an LLM for generation of response.

Krissy 2 years ago

I think option A is the best choice because it allows the LLM to summarize news articles and find stock prices efficiently.

Databricks Certified Generative AI Engineer Associate Exam - Topic 3 Question 2 Discussion