Passing the Google Data Analytics Interview

Will AI pass the data analytics technical interview?
Gemini
Data Analysis
Communication
Author

Jason Saito

Published

March 19, 2024

Introduction

I wanted to see if various LLMs could pass the technical side of the data analytics interview. I first began by running the model through Google’s interview simulation.

Technical Question #1

What are the advantages of using a version control system like Git?

Reply

image.png Overall, the model performed well and was able to satisfy Google’s three main answer criteria of job-related terms, most-used words, and talking points.

Situtational Question #1

I then asked it a situational question.

Reply

image.png

It seems that this response is much weaker than the technical answer before.

Prompt #1

I first provided the model with an easy leetcode problem about databases.

image.png

Response from Gemini Pro

The response was accepted on the first try with a pretty average efficiency and runtime.

Prompt #2

I then provided the Gemini Pro model with a medium leetcode problem.

image.png

Response

This response was also accepted on the first try with pretty average efficiency and runtime.

Prompt #3

I then provided a hard leetcode problem.

image.png # Response The model got the question wrong three times in a row, even after I provided it the feedback from leetcode on what test case failed.

Conclusion

Overall, Gemini Pro did pretty well against both the interview questions and the coding questions. It seems to struggle with more personalized questions and advanced data analytics programming questions.