← Back to Directory

Uber interview notes, page 11 of 61

Large Enterprises

Most of these Uber write-ups come from Software Engineers (407), then Data Scientists (78) and Machine Learning Engineers (65). The most reported rounds are the technical phone screen (387), the system design round (95), the product sense round (46) and the onsite coding round (23).

609 interview notes · updated Aug 2026

Uber·Machine Learning Engineer·Onsite - Multi Round

Apr 2026
Uber MLE loop with two parts: a deep ML system retrospective and a behavioral block focused almost entirely on conflict and disagreement. The system design piece was more open-ended than I expected, and the behavioral round felt unusually pointed about how you handle pushback.
  • Walk through an ML system you designed or heavily contributed to, covering problem framing, data, modeling decisions, online vs offline design, evaluation, and what you'd change now.
  • Tell me about a direct conflict with a teammate over a technical decision. How did you approach the conversation and what happened?
  • Describe a time you disagreed with your manager or a senior stakeholder. What data or framing did you use, and how was it resolved?
  • Tell me about a persistent disagreement you had to either fully resolve or ultimately commit to despite not agreeing. What made you decide which path to take?

“This one took up a lot of time and I don't think I paced it well.”

View Post

Uber·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Uber MLE interview with a single coding question that was more math-flavored than I expected. No LeetCode number, just a black-box optimization problem. Felt pretty clean once I landed on the right approach, though I almost overthought the termination condition.
  • You're given a black-box convex function F(x) that you can only evaluate by querying it at specific x values. Given a search interval [a, b], find the x that minimizes F(x). You cannot compute gradients, only query function values.

“Ternary search clicked for me pretty fast since the function is unimodal.”

View Post

Uber·Software Engineer·Technical Phone Screen

Apr 2026
Uber SWE coding round, pretty standard stuff. Got an LRU cache problem which I'd seen before but still fumbled parts of the implementation under pressure.
  • Design and implement an LRU cache with get and put operations, both running in O(1) average time complexity.

“I knew the answer going in: hash map plus doubly linked list.”

View Post

Uber·Product Manager·Onsite - Product Sense / Strategy

Apr 2026
Uber PM interview with a metrics deep-dive question. Pretty standard product sense round but the conversion rate angle made it more analytical than I expected.
  • If you noticed a 10% drop in an app's conversion rate, how would you investigate and fix it?

“I jumped straight into segmentation before even asking what 'conversion' meant in this context, which was probably the wrong move.”

View Post

Uber·Software Engineer·Technical Phone Screen

Apr 2026
Uber SWE interview with a twist on a classic problem. Pretty short on details but the question itself is worth thinking through.
  • Given a stream of IP addresses hitting a server, find the first IP address that appears exactly once.

“Basically a 'first unique character' problem but with IPs instead of chars.”

View Post

Uber·Software Engineer·Technical Phone Screen

Apr 2026
Uber SWE interview with a graph problem that looked manageable until I actually had to implement it cleanly under pressure. The optimal path requires sorting both queries and edges, then running union-find incrementally, and there are enough moving parts that small bugs will kill you.
  • Given an undirected weighted graph and a list of queries each specifying two nodes and a weight limit, determine for each query whether a path exists between the two nodes using only edges with weight strictly less than the given limit.

“The naive approach of running a fresh BFS per query will time out.”

View Post

Uber·Software Engineer·Technical Phone Screen

Apr 2026
Uber SWE coding round, just one algorithmic problem the whole session. Pretty focused interview, nothing behavioral at all.
  • Given an array of prices, for each item find the first later item whose price is less than or equal to it and use that as a discount. Return the array of final discounted prices.

“Classic monotonic stack problem once you see it, but I spent a few minutes trying to brute force it first.”

View Post

Uber·Software Engineer·Technical Phone Screen

Apr 2026
Uber SWE interview with a permutation problem that looked straightforward on the surface but had some real edge cases hiding underneath. The expected complexity was O(n log n) or better, so brute force was a non-starter.
  • Given a permutation of integers 1 through n, for each k from 1 to n determine whether there exists a contiguous subarray of length k that contains exactly the values 1 through k. Return a binary string of length n where position k-1 is '1' if such a subarray exists and '0' otherwise.

“My first instinct was a sliding window but the window size changes with k so that doesn't cleanly apply.”

View Post

Uber·Software Engineer·Online Assessment (OA)

Apr 2026
Uber SWE online assessment, one algorithmic problem via Hack2Hire. Pretty standard sliding window stuff but the edge cases can bite you if you're not careful.
  • Given an integer array and a value k, find the length of the shortest contiguous subarray containing at least k distinct integers. Return -1 if no such subarray exists.

“Sliding window with a frequency map, not too bad once you see it.”

View Post

Uber·Software Engineer·Onsite - System Design / Architecture

Apr 2026
Uber system design round focused entirely on building a distributed logging system for a massive microservices platform. Multi-part question that went deep into ingestion, storage tiering, and operational concerns. Pretty grueling but the structure helped.
  • Design a distributed logging system for a large microservices platform handling millions of log events per second across multiple regions, from ingestion all the way to searchable storage.
  • How does a log line travel from an application process to a durable buffer, and what happens on the host agent side in terms of batching and back-pressure?
  • How would you design the storage and indexing layer, including how you tier logs from hot searchable storage to cheap long-term archival?
  • A single high-cardinality field like a unique request URL with embedded IDs is blowing up your index size. How do you detect this and what do you do about it?
  • An engineer needs to trace a single request across 30 services. What needs to be in the logging schema and transport path to make that possible, and how does it interact with sampling?
  • A region goes fully offline for 2 hours and comes back. What happens to logs produced during the outage and how do they become searchable afterward?
  • The business wants to cut logging costs by 40% without losing the ability to debug production incidents. What levers do you pull and how do you verify debugging capability is preserved?
  • How do you prevent one noisy service from starving other tenants in a shared logging pipeline?
  • How do you monitor the logging pipeline itself without depending on the very index it feeds?

“This is the core question and it's massive.”

View Post