Data Scientist · Phone Screen · Multiple Types

Intuit — Data Scientist ✅ Passed

Level: Senior-Level

Round: Phone Screen · Type: Multiple Types · Difficulty: 6/10 · Duration: 60 min · Interviewer: Friendly

Topics: SQL, Python, Product Sense, Classification Models, Precision, Recall, F1 Score, Data Analysis, Machine Learning

Location: San Francisco Bay Area

Interview date: 2026-02-15

Summary

Round 1: Phone Screen

Question: SQL: Calculate cohort-based retention rate.

Details

This was a technical phone screen for a Product Data Scientist role. However, the interviewer was from the AI team, so the questions might not have been typical.

I was asked:

• SQL: Calculate cohort-based retention rate (classic problem).
• Python: Given a set of classification model outputs (e.g., [actual =1, predict = 0, conf = 0.93]), calculate precision, recall, and F1 score. I was told I could use pseudocode if I wasn't comfortable with Python.
• Product Sense:
  1. Given user address information with zip codes, what data can be associated as input variables for a prediction model? I suggested using state and city, or census data like household income and education level.
  2. If zip codes are missing, how would you handle it? I said if there were very few missing, I would remove them. Otherwise, I would find the most common zip code for that address or create a new variable called without zip code