
Instacart Data Scientist Interview Questions
& Process
Real candidates share what happened, how many rounds they had,
and how the experience turned out.
Based on 34 interview experiences · FREE TO READ
Candidate interview experiences
First-hand accounts from people who interviewed at Instacart.
Data Scientist Intern
I had a 45-minute technical interview focused on statistics and SQL, with questions tailored to Instacart scenarios. The SQL part wasn't about writing code from scratch, but more about spotting mistakes in existing SQL queries. It seems like companies are shifting their interview strategies, perhaps because LLMs can generate code so well now, and the focus is moving towards evaluating the ability to identify flawed LLM outputs. The interviewer seemed pretty laid-back from the beginning and didn't make much effort to create a welcoming atmosphere.
No confirmed questions were included in this interview report.
Data Scientist
The interview process started with a tech screening, then a take-home assessment where I had to review an experiment and provide a recommendation. After that, there were four rounds for the virtual on-site. One of these rounds was to go over the assessment I had previously submitted, and the others were standard interviews, including one with a bar raiser and the hiring manager.
- Can you discuss a past project in detail?
- Describe a project where you reviewed an experiment and provided a recommendation.
- Please present your findings and recommendations from the take-home assessment.
Data Scientist
The interview process was straightforward, starting with HR and then moving to technical interviews. The technical part included questions on statistics and math. Everything moved quite fast. My interviewer was very late without an apology and generally unfriendly, but a friend interviewing at the same time had a better experience and got the same questions.
- If we were to introduce 15-minute deliveries in an existing delivery area, what statistical methods would you use to evaluate the outcome?
Instacart Data Scientist Interview Questions
Quoted word for word from Instacart interview reports.
“Regarding right-skewed distributions, is it appropriate to apply a log transform?”
Read reports →“What are the steps for tuning a random forest model?”
Read reports →“Is it acceptable to use retention rates as a measure of user satisfaction?”
Read reports →“Can you compare and contrast random forest with gradient boosting?”
Read reports →“For an A/B test analysis on the thanksgiving-coupon experiment, where users are either in the incentive group (getting $10 off) or non-incentive group, and the coupon expires on 11/29/2020 PST, how would you analyze the results? You have access to the Experiments table (user_id, experiment, group_ind, assigned_at_utc) and the Orders table (user_id, order_id, order_created_timestamp_utc, order_amount_usd), and you can convert UTC timestamps to PST. You also have the Coupon_spend table (user_id, order_id) for tracking redeemed coupons.”
Read report →“Can you explain how to perform power analysis for an A/B test?”
Read report →“Describe your approach to using logistic regression for predicting a continuous outcome.”
Read report →“Explain how you would utilize logistic regression to forecast the amount of a purchase.”
Read report →“If we were to introduce 15-minute deliveries in an existing delivery area, what statistical methods would you use to evaluate the outcome?”
Read report →Formats, difficulty and experience
Across all 34 Instacart interview reports.