LinkedIn Interview Questions — Page 9

LinkedIn·Data Scientist·Technical Phone Screen

May 2026

LinkedIn data scientist interview with a meaty causal inference question about a mobile app redesign rollout. No forced experiment, so you have to reason your way through observational methods. Solid technical depth required.

A mobile app is redesigned but users self-select into the new version. Without running a forced experiment, how would you estimate the causal impact of the redesign? How do you define treatment and control groups, which user features would you use to ensure comparability, and what methods would you apply to validate your assumptions?

“This one is where I started okay and then slowly dug myself a hole.”

View Post

LinkedIn·Software Engineer·Technical Phone Screen

May 2026

LinkedIn SWE screen, pretty standard stuff. They opened with a classic string reversal problem, in-place, O(1) memory. Nothing unexpected.

Reverse a string in-place given an array of characters, using only O(1) extra memory.

“Two pointers from both ends, swap until they meet.”

View Post

LinkedIn·Software Engineer·Technical Phone Screen

May 2026

LinkedIn SWE interview that was basically a harder version of the classic insert/delete/getRandom problem, except they bolted on TTL semantics and wanted a real design discussion. Pretty gnarly for a coding round.

Design a data structure that supports insert(value, ttl), remove(value), and getRandom() in O(1) amortized time, where each inserted item has a time-to-live after which it should be treated as logically removed. Discuss how you'd handle duplicates of the same value with different TTLs, and compare lazy expiration against active eviction strategies like a priority queue or timer wheel.

“So this is LC 381 but with TTL thrown on top, which sounds manageable until you actually try to keep getRandom truly O(1).”

View Post

LinkedIn·Product Manager·Onsite - Product Sense / Strategy

May 2026

LinkedIn PM interview, got asked about evaluating Gen AI systems. Single question, not a lot of context around it but it's the kind of thing that trips you up if you haven't thought about it before.

How would you evaluate the quality and effectiveness of a generative AI system?

“I fumbled the structure a bit at first.”

View Post

LinkedIn·Product Manager·Hiring Manager Screen

May 2026

Interviewed for a PM role at LinkedIn and got hit with a question about Responsible AI and regulatory frameworks. Pretty niche for a PM screen but I guess it makes sense given the platform.

What do you know about Responsible AI principles and the regulations that currently govern how AI can be used?

“I fumbled around with some general stuff about fairness and transparency and then tried to name-drop a regulation or two.”

View Post

LinkedIn·Product Manager·Technical Phone Screen

May 2026

LinkedIn PM interview with one question about endpoints. Pretty light on content so hard to say much about the overall process.

What is an endpoint? Give two examples.

“Straightforward enough but I second-guessed myself on whether they wanted a networking definition or something more API-specific.”

View Post

LinkedIn·Data Scientist·Technical Phone Screen

May 2026

LinkedIn Data Scientist whiteboard round, heavy on math. Four parts back to back: logistic regression from scratch, backprop derivations, mini-batch pseudocode, and Adam internals. Not a conceptual chat at all, they wanted chain-rule steps on the board.

Walk through logistic regression for binary classification: write the model, the sigmoid, and the binary cross-entropy loss for a single example and for a batch, then derive the gradient with respect to the weights and bias from first principles.
Show that logistic regression is a special case of a neural network, then write the full forward pass and backpropagation equations for an L-layer feedforward network, including how gradients flow through each layer.
Write pseudocode for mini-batch gradient descent training and explain why mini-batches are used instead of full-batch gradient descent or single-example updates.
Describe Adam's update rule, write out all four components including bias correction, and give a real argument for why the second moment estimate appears under a square root in the denominator.
Why is binary cross-entropy preferred over mean squared error for classification with a sigmoid output? What goes wrong with the optimization landscape if you use MSE instead?
What is the vanishing gradient problem in networks with sigmoid or tanh activations, and what approaches address it?
How does Adam compare to RMSProp and to SGD with momentum, and are there situations where plain SGD with a learning rate schedule generalizes better?
What is weight decay, and why is decoupled weight decay (as in AdamW) different from just adding an L2 penalty to the loss when training with Adam?

“The derivation itself isn't hard but I kept second-guessing whether to show the sigmoid derivative cancellation explicitly or just state it.”

View Post

LinkedIn·Software Engineer·Technical Phone Screen

May 2026

LinkedIn coding round for a software engineer role. One problem, pretty involved, centered on graph traversal with some file parsing thrown in. Felt like a solid mid-level systems-thinking question dressed up as a coding exercise.

You're given a JSON file describing services and their dependencies. A specific service goes down. Write code to parse the file, build a dependency graph, and find all services transitively affected. Be ready to talk about time complexity and how you'd handle cycles.

“The JSON parsing part was fine but I fumbled a bit explaining why I chose BFS over DFS.”

View Post

LinkedIn·AI Engineer·Onsite - Behavioral / Leadership

May 2026

Behavioral round at LinkedIn for an AI Engineer role. Just one question but it had a lot of moving parts and I wasn't totally prepared for how deep they wanted to go on the influence-without-authority angle.

Tell me about a time you influenced a design decision when you didn't have direct authority over it. What was the original direction, what did you push for instead, how did you make your case, and what happened?

“The part that tripped me up was the 'how did you build the case' section.”

View Post

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

May 2026

LinkedIn MLE system design round focused entirely on the Skills platform, which sounds scoped until you realize they want you to cover inference, recommendation, serving, evaluation, and abuse handling all in one go. Pretty demanding for a single session.

Design an ML system for LinkedIn Skills that can infer skills a member has, recommend new ones, and serve downstream use cases like search ranking, recruiter matching, and personalization.
What data sources would you use to label and train a skill inference model, and how would you handle the labeling strategy?
How would you use LLMs in the feature engineering or modeling pipeline for skill inference or recommendation?
Walk through how you would architect the online serving layer for skill recommendations, including latency and freshness considerations.
How would you evaluate the skill inference and recommendation system both offline and online, and how would you detect and handle quality issues like hallucinated skills, stale skills, or spam?

“This is a beast of a question.”

View Post