← Back to Directory

Pinterest interview notes, page 13 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·Machine Learning Engineer·Technical Phone Screen

Jun 2026
Pinterest ML Engineer interview, open-ended technical discussion round focused on multimodal models. The interviewer was willing to pull up a paper live and talk through it together, which I wasn't expecting at all.
  • Walk me through how modern multimodal models combine vision and language, including the architecture components and training stages involved.
  • Describe the architecture, training data, pretraining objectives, and key design decisions in the Qwen-VL paper.
  • How would you extend or evaluate a multimodal vision-language model like Qwen-VL?
  • Given an unfamiliar section of a research paper, can you read and reason through it on the spot?

“This is broader than it sounds.”

View Post

Pinterest·Software Engineer·Onsite - System Design / Architecture

Jun 2026
Pinterest system design round for a software engineer role. The whole session was basically one giant question about building distributed blob storage, and they went deep on every layer. Not a casual chat.
  • Design a globally distributed object storage service similar to Amazon S3, covering core operations like storing and retrieving objects, large file uploads, versioning, and access control.
  • How would you approach durability guarantees at massive scale, and what are the trade-offs between simple replication and erasure coding?
  • Walk through the full data path from a client upload request to data being durably stored across multiple availability zones.
  • How would you handle garbage collection and repair in a distributed storage system where objects can be deleted or partially written?
  • How would you design the metadata service to stay consistent under high load, and what happens when a single bucket or key becomes a hotspot?

“This was the whole interview.”

View Post

Pinterest·Software Engineer·Technical Phone Screen

Jun 2026
Pinterest coding round for a software engineer role. One problem, fairly focused on greedy balance settling. Not the hardest problem I've seen but the 'to' field being a list tripped me up at first.
  • Given a list of transactions where each transaction has a sender, one or more receivers (amount split equally), and a total amount, return any list of paybacks that brings every user's net balance to zero. You don't need to minimize the number of paybacks.

“The multi-receiver part is what got me initially.”

View Post

Pinterest·Data Scientist·Technical Phone Screen

Jun 2026
Pinterest data scientist interview with an algorithms problem dressed up as a product scenario. The debt-settlement framing was a nice touch but the core of it is just a graph/greedy problem once you strip it back.
  • You're building an expense-sharing app. Given a list of transactions where each entry has a payer, a payee, and an amount, design an algorithm that finds the minimum number of payments needed to settle all balances among the group.

“I spent the first couple minutes thinking about this as a graph problem and almost went down a max-flow rabbit hole.”

View Post

Pinterest·Data Scientist·Technical Phone Screen

Jun 2026
Pinterest coding screen for a Data Scientist role, and they threw a compression/decoding problem at me that felt more like a software engineering interview than anything data-related. The ambiguity angle was the part that actually tripped me up.
  • Implement a decode function for a compression format where strings like '1234' encode a sequence of values by count-value pairs (e.g. '1234' becomes [2,4,4,4]). Then explain why an input like '12114' is ambiguous and enumerate all valid decodings.

“The basic decode wasn't too bad once I realized you parse a multi-digit count followed by a single digit value.”

View Post

Pinterest·Machine Learning Engineer·Technical Phone Screen

Jun 2026
Pinterest ML engineer screen, just one technical question about activation functions. Pretty standard stuff but I fumbled the depth a bit.
  • What are the key differences between ReLU and Sigmoid activation functions?

“I covered the basics fine, output ranges, the vanishing gradient problem with sigmoid, ReLU's sparsity.”

View Post

Pinterest·Machine Learning Engineer·Technical Phone Screen

Jun 2026
Pinterest ML engineer screen, one multiple-choice style question on ensemble methods. Pretty focused, felt more like a knowledge check than a deep technical discussion.
  • What are the trade-offs between bagging, boosting, and stacking as ensemble strategies? When does each approach help or hurt?

“Multiple choice framing threw me off a bit because I kept second-guessing whether they wanted the textbook answer or something more nuanced.”

View Post

Pinterest·Machine Learning Engineer·Technical Phone Screen

Jun 2026
Pinterest MLE interview that came down to a graph problem I hadn't touched in a while. The algorithmic depth required here was a bit more than I expected for an MLE role.
  • Given a list of airline tickets where each ticket is a [departure, arrival] pair, reconstruct a valid itinerary starting from JFK that uses all tickets exactly once. If multiple valid itineraries exist, return the one with the smallest lexical order.

“I knew this was an Eulerian path problem pretty quickly but then froze on the implementation.”

View Post

Pinterest·Machine Learning Engineer·Technical Phone Screen

Jun 2026
Pinterest ML engineer interview had at least one multiple-choice question on confusion matrices, which sounds easy until you're second-guessing which axis is which under pressure.
  • Given a confusion matrix, compute precision, recall, F1, accuracy, and specificity. Also interpret the matrix correctly given the axis labeling shown.

“Multiple choice format, which made it feel easier than it was.”

View Post

Pinterest·Data Scientist·Technical Phone Screen

Jun 2026
Pinterest data scientist round, basically a coding problem about computing free time slots in a restaurant booking system. More algorithmic than I expected for a DS role, felt like it belonged in a software eng loop.
  • Given a restaurant's opening hours and a list of existing reservations per day, write a function that returns the available time intervals for each day. Reservations may be unsorted or overlapping and should be merged before computing free slots.

“My first instinct was to just iterate and subtract, which works until you have overlapping reservations and suddenly your logic is wrong.”

View Post