🚖 Uber SDE-2 Interview Experience (Bangalore)

Process: Referral
My Experience: 3 years as SDE-1
Total Rounds: 5 (1 elimination coding + 4 interviews)

📌 Round 0 – Elimination Coding Round

• Problem Statement:
  Given an array of integers nums and an integer limit, return the size of the longest non-empty subarray such that the absolute difference between any two elements of this subarray is ≤ limit.
  (Similar to LeetCode 1438 – Longest Continuous Subarray With Absolute Diff ≤ Limit)

• Approach:
  The question was framed in a driver-cab domain context. I explained four approaches step by step:

  1. Brute force (4 loops) → Generate all subarrays and check conditions.
  2. Improved brute force (3 loops) → Avoid redundant calculations.
  3. Further optimization (2 loops) → Track min/max with sliding windows.
  4. Optimal solution (Deque + Sliding Window, O(n)) → Maintain min/max using two deques.

  The interviewer was very particular about base cases and explaining time & space complexity for each approach.

✅ Cleared and moved to next rounds.

📌 Round 1 – DSA + DSU Problem

• Problem:
  We were given logs in the format UserA-Shared-UserB-timestamp, meaning UserA and UserB share a cab.

  • Version 1: For each log, return how many cabs are currently running.
  • Version 2: For a given logs , return the timestamp when all users get into one cab.

• My Approach:

  • Started with DFS/BFS for connected components.
  • Shifted to DSU (Disjoint Set Union) for optimal handling.
  • Implemented DSU with path compression + union by rank/size.
  • Interviewer asked me to code DSU from scratch and explain parent finding, rank-wise clustering, and size-based union.
  • Explained time and SPace Complexity for the same

📌 Round 2 – Low-Level Design (LLD)

• Problem Statement:
  Design a Train-Platform Management System with functionalities:

  • Assign trains to platforms based on input.
  • Query which train is at a given platform at a specific time.
  • Query which platform a train is at, at a specific time.

• Expectations:

  • Write runnable code (in my preferred language – chose Java).
  • Explain design patterns used.
  • Walkthrough every step while coding.

• My Approach:

  • Designed classes: Train, Platform, Scheduler, Schedule Manager.
  • coded using minHeap Startegy and then explained random Assigning Startegy , wrote extensible code
  • Focused on time-based queries and mapping train schedules.
  • Did not added all class variables , just explained and just added variables requried for getting answer ( for example , for trian I jus added trian id , not capacity or type )
  • The interviewer continuously cross-questioned choices and pushed for clean design + working code in 60–75 minutes.

📌 Round 3 – High-Level Design (HLD)

• Problem Statement:
  Design Uber Eats Homepage.

• Focus Areas:

  • Location-based search → GeoHashing & GeoIndexing.

  • API design for homepage and search results.

  • Schema design for Restaurants, Dishes, Orders.

  • Choice of database (SQL vs NoSQL) and justification.

  • Use of caching for performance.

  • Metrics to capture:

    1. Most ordered restaurant.
    2. Most ordered dish.
    3. Orders per restaurant.

• Tools: HackerRank’s design whiteboard/drawing tool to draw architecture.

• Interveiwer was so frendly and was listeinig and supporting all approches and at the same time pointing out the mistakes in my explanation as well

📌 Round 4 – Managerial + Behavioural

• Discussion around past work experience:

  • Good coding practices followed.
  • My role in previous projects.
  • Ownership & impact of work delivered.

• The manager also gave me ~15 minutes to ask questions about the team, role, and growth opportunities.

✅ Outcome

Finally, I received the offer for Uber SDE-2, Bangalore 🎉