Hi all, I had my onsite loop with Pinterest last month for their Sr. MLE position. I did not get the offer, but I still wanted to share my experience here, so it can benefit others.

Timeline

Screening round - Aug last week - 1 hour single round

Onsite round - split over 2 days - mid-September

• Day 1 - Coding
  • 2 rounds
• Day 2
  • HM interview - behavioral
  • ML practitioner
  • ML system design

Rejection email - one week after last onsite round

Screen Coding round

Asked project explainer for 10 mins,
3 ML questions for 10 mins (quick round),
Then 2 coding questions related to each other + bonus coding question (also related)

• DFS question (with 2 parts)
• Optional Bonus question - follow up that had a DP solution (only idea discussed, not asked to code)

Onsite round

Coding #1

• I was asked to implement a data structure that I knew about but had not practiced before (not on Pinterest or Meta recent questions list)
• I struggled in this round, not because I did not know what needed done, but it was my first onsite and I bombed out of nervousness

Coding #2

• Simple twist over a top-100 question from another company (Meta). Also easy to do even if you have not seen the question before. Array type question.
• I did well here. Also solved+coded an optional follow-up

HM interview

• HM was seasoned, so he did not just give me the floor and repeat my practiced monologue for my experience. I think I did okay, but came out less imnpressive than I have done in other HM interviews.
• overall, the focus includes the usual senior stuff i.e., cross functional communication, mentoring, one technical depth type story, conflict resolution. But I noticed that unlike big tech companies, Pinterest seemed to care about engineering related details as well. For example, at one point, I was asked what my tech stack is - this, I was unprepared for, but I managed to give some answer - he did not seem too impressed with what I said.

ML practioner round

• This round took me by surprise. I had expected this to be another design round. Maybe it was some miscommunication between the recruiter and me, but I went unprepared into this round, given what was asked. This was a 2-part interview.
• In the first part, the interviewer wanted to discuss any of the research articles of mine. I had been working in the industry for 5 years and most of my papers were from graduate school 5+ years ago. I told them that, to communicate that discussing something from 5+ years ago is a bit irrelavent as far as my industry experience goes. But they seemed to still want to discuss it which I though was odd, but we I picked a paper that they seemed to like and walked them through the main idea/architecture of the solution. They seemed to follow and be satisfied with the answers.
• They wanted to know, the problem statement, intuition of the solution, model architecture and some details on baseline methods implemented in the study.
• In the second part of the interview, they asked me details on ML topics. Some of the questions were more specific than what would be discussed in a ML course.
• For example, I was asked to explain the transformer architecture, which is standard these days. But I was also asked to explain constrastive loss, which I knew but I thought was non-standard. Then there were questions on learning to rank algorithms, which I knew only surface-level details.
• Overall, I think I did above average in this round, but I missed 1-2 quesitons on the ML topics.

ML system design

This was a standard ML system design question. I recommend people do a standard set of these questions including all details therein. I followed this - https://bytebytego.com/intro/machine-learning-system-design-interview

Rather than doing more case studies, I recommend understanding all case studies to a good level of depth. Every detail you say in the design, you should know details around. For example, I did not know contrsative learning until I did the similarity search type case studies in the ML design resource.

Outro

Overall, I found the depth of interviewers at Pinterest to be higher than the 2 other onsites I gave (one FAANG and another medium sized company with FAANG-pay).

The manager and the other interviewers were also nice to interact with, so I assume it is a good place to work.

I did not receive explicit feedback, but I think the first coding round tanked my case.

Good luck, folks! I hope this helps folks prepare for their interviews!