Google Data Scientist Interview Experience — An Unexpected Forecasting Question and a Fast Offer Timeline

I signed Google's offer a couple of days ago, though I still have a few other issues to resolve. I'm sharing my interview experience and preparation package, and hoping I can start smoothly.

First round
Stats: Questions about p-values, simple linear regression, confidence intervals, and similar topics.
Data: Compare two vendors' models. There are similar questions in other interview reports.

Second round
Coding: Basic data manipulation, with some discussion of how to choose metrics or methods.
Data: The interviewer spontaneously came up with a forecasting question I'd never seen. It mainly tested modeling intuition.
Behavioral: Standard questions.

The overall experience was very good. The interviewers were friendly and didn't nitpick extremely fine details; it mainly felt like a discussion. The second-round data question was particularly unexpected, but the interviewer was very nice and we had a really enjoyable conversation.

Timeline: The recruiter reached out in mid-July, and we immediately scheduled an interview for a few days later. About three days after the interview, I heard that I was moving forward. This was around late July. I quickly scheduled an early-August onsite, split over two days. Roughly three days afterward, in mid-August, I heard that the feedback was positive and I could start team matching. I matched about a week later. We negotiated compensation, and another week later, in early September, I received the offer. I was eager to start, so everything moved very fast. Still, moving forward within three to five business days after each step seems about right.

I've attached some preparation notes and sample questions I put together, mostly from questions I found in other interview reports and summarized myself. The DSR interviews barely tested the product material. I haven't interviewed for DSP, but I can confirm from my own interviews that Meta and TikTok ask it. I feel I overprepared the complexity of the stats and coding material; use whatever you need.

Product notes, statistics notes, and R coding notes.