← Back to Directory

Pinterest interview notes, page 3 of 24

Large Enterprises

Most of these Pinterest write-ups come from Machine Learning Engineers (91), then Software Engineers (77) and Data Scientists (51). The most reported rounds are the technical phone screen (168), the system design round (37), the product sense round (14) and the behavioral / leadership round (5).

235 interview notes · updated Sep 2026

Pinterest·Software Engineer·Technical Phone Screen

Apr 2026
Pinterest coding interview with a classic grid traversal problem. Nothing too surprising but it still required clean execution under pressure.
  • Given a 2D binary grid representing a map where '1' is land and '0' is water, count the number of islands. An island is a group of adjacent land cells connected horizontally or vertically.

“Classic BFS/DFS problem and I knew it immediately, which almost made it worse because I rushed into code without talking through edge cases first.”

View Post

Pinterest·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Pinterest MLE coding screen focused entirely on string manipulation for numeric rounding, and the twist was you couldn't use any built-in float parsing at all. Two problems, both variations on the same theme but the second one added a precision parameter that made things messier than expected.
  • Implement a rounding function from scratch that takes a decimal number as a string and returns the rounded integer as a string, without parsing it into a float.
  • Extend the rounding function to round a numeric string to the nearest multiple of a given precision, where both the number and the precision are provided as strings.

“The no-float constraint is the whole point of the problem and it took me a minute to internalize why.”

View Post

Pinterest·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Pinterest ML Engineer round focused entirely on log analysis, two parts back to back. Pretty algorithmic for an MLE role but not unreasonable once you see where it's going.
  • Given a list of violation log entries each containing a user ID, timestamp, and violation type, design a data structure to aggregate these logs and implement two methods: one that returns the violation count for a given user, and one that returns all users whose violation count exceeds a threshold k.
  • As a follow-up to the log aggregation problem, implement a method that returns the latest violation timestamp for a given user that is strictly before a given time T, and do it efficiently.

“The aggregation part was fine, hash map keyed by user ID storing a list of timestamped entries.”

View Post

Pinterest·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Pinterest MLE interview with a classic group expense balancing problem. Pretty much a direct lift from LC 465 so if you've seen it before you're in decent shape, but the backtracking angle can trip you up if you're not warmed up on DFS.
  • Given a group trip where each person paid different amounts for shared expenses, find the minimum number of transactions needed to settle all debts among the group.

“Knew this one from grinding leetcode but still fumbled the opening explanation.”

View Post

Pinterest·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Pinterest ML Engineer interview with a warehouse packing problem that looks straightforward until you actually try to code it under pressure. One question, greedy approach, and a lot of second-guessing myself in the moment.
  • You have an array of box heights and an array of warehouse room heights (left to right). Boxes can only enter from the left, and a box can only reach room j if its height is at most the minimum height of all rooms from 0 to j. Each box takes exactly one room. What is the maximum number of boxes you can fit?

“My first instinct was to sort both arrays and just greedily match, but the constraint that boxes push through all prior rooms tripped me up.”

View Post

Pinterest·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Pinterest ML Engineer interview touching on a classic model behavior question. Not a lot of context to go on but the core topic was overfitting and the gap between training and test performance.
  • Your model has high training accuracy but much lower test accuracy. What's going on and how do you fix it?

“Classic overfitting question but it's easy to ramble here.”

View Post

Pinterest·Software Engineer·Onsite - System Design / Architecture

Apr 2026
Pinterest system design round focused on sparse matrix storage and operations. Pretty niche problem, not the usual distributed systems stuff I prepped for.
  • Design a data structure for storing a sparse matrix efficiently, and implement addition and multiplication operations on two such matrices without wasting computation on zero elements.

“I went with a hash map keyed on (row, col) pairs storing only non-zero values, which felt reasonable.”

View Post

Pinterest·Machine Learning Engineer·Onsite - System Design / Architecture

Apr 2026
Pinterest ML engineer interview focused heavily on recommendation systems, specifically around multi-vector user representations. The depth expected was pretty serious, covering everything from clustering approaches to serving infrastructure and evaluation.
  • Given pre-trained item embeddings, how would you design a system to represent each user with multiple embeddings instead of a single averaged vector?
  • How would you decide whether to use a fixed number of embedding heads per user versus letting that number vary adaptively?
  • What training signals would you use to learn these multi-vector user representations, and how do you handle negatives?
  • Once you have multi-vector user embeddings at serving time, how do you actually use them to retrieve candidates efficiently?
  • How would you evaluate whether multi-vector user embeddings actually improve things, both offline and in a live product?

“This is a meaty one.”

View Post

Pinterest·Machine Learning Engineer·Onsite - System Design / Architecture

Apr 2026
Pinterest system design question focused on recommender systems. Pretty technical and specific, the kind of thing you either know from having built it or you're doing math on the fly.
  • In a two-stage recommender system with retrieval and ranking, how do you decide how many candidates to pass from retrieval into the ranking model? Walk through the tradeoffs around inference cost, latency, and cluster capacity.

“This one required actual back-of-the-envelope thinking.”

View Post

Pinterest·Software Engineer·Technical Phone Screen

Apr 2026
Pinterest coding screen for a software engineer role. Just the one problem but it was a bit of a weird one if you haven't seen it before.
  • Implement the count-and-say sequence: given a positive integer n, return the nth term as a string, where each term is built by run-length encoding the previous term.

“I knew what run-length encoding was but hadn't connected it to this problem before.”

View Post