Uber Freight SWE3 Interview Experience
Anonymous User
357
UberCareerFeedbackInterview

Uber Freight SWE 3 Interview Experience

Here’s my interview experience for the Uber Freight SWE 3 role.

Round 1 — DSA

We were given a car and N riders.

For every rider, we have two values:

  • L[i] — minimum number of other riders this rider is willing to travel with.
  • H[i] — maximum number of other riders this rider is willing to travel with.

If we select k riders, then every selected rider must satisfy:

L[i] <= k - 1 <= H[i]

The goal is to maximize the number of riders that can travel together.

Example:

N = 5

L = [0, 1, 1, 2, 2]
H = [1, 2, 2, 4, 4]

Answer = 3

The expected solution was O(N).

The main challenge was identifying the right way to think about the constraints and avoid checking every possible k.

Round 2 — LLD

Design a basic file system supporting operations such as:

  • mkdir
  • ls

The interviewer was mainly interested in the design and approach, rather than implementation.

We discussed:

  • File and directory entities
  • Directory hierarchy
  • How to represent parent-child relationships
  • How mkdir would create a directory
  • How ls would traverse/list the contents
  • Entity/class relationships

We were expected to draw the entity/class diagram and explain the design.

Working code was not expected.

Round 3 — System Design

Design a distributed URL shortener.

The discussion covered the overall distributed architecture and how the system would work at scale.

Important areas to consider:

  • URL shortening API
  • Redirect API
  • Unique short-code generation
  • Database/storage
  • Caching
  • Distributed ID generation
  • Read/write scalability
  • Handling very high redirect traffic
  • Availability and fault tolerance
  • Expiration of URLs
  • Collision handling
  • Database partitioning/sharding

Overall, the rounds were a mix of DSA → LLD → HLD, with increasing emphasis on design depth and scalability.

Difficulty: DSA was mainly about finding the optimal observation, LLD focused on clean object modeling, and the final round tested distributed-system fundamentals and scalability.

I wanted to discuss on compensation part, do you guys any idea of how much they offer?

Comments (2)