All experiences
Offered2026-08-25·3 min read

Uber SDE-II (L4) Bangalore — 5 Rounds, 3 In-Person

L4 loop at Uber Bangalore. First 3 rounds in-person (screening, DSA, LLD), then virtual HM and HLD. Graphs are Uber's favorite. Whole process ~2 months. Referral-based.

Company

Uber

Role

SDE-II (L4)

Rounds

5

YOE

4

Timeline

  • Referral → recruiter reached out
  • 5 rounds — first 3 in-person, last 2 virtual
  • Entire process: ~2 months
  • Offer: ~3 weeks after final round

Background

  • 3.7 YOE
  • Tier-1 education, currently at a Tier-1 company
  • Interviewed April 2026 for L4 (SDE-II)

Round 1 — BPS / Screening (In-Person) — Strong Hire

Problem: Robot navigating a grid with charging cells. Find optimal path with priority: minimum charging cells → minimum battery required → minimum moves.

Difficulty: Hard

This was a multi-criteria state-space search (modified Dijkstra with a tuple cost). The interviewer wanted:

  • Logic overview with a dry run first
  • Clean, optimal running solution
  • Follow-ups

Being in-person changed the dynamic — more collaborative, whiteboard + laptop.

Round 2 — Coding 1 / DSA (In-Person) — Strong Hire

Problem: Find Median from Data Stream

Difficulty: Hard

Expected approach:

  • Start with suboptimal (sorting / insertion)
  • Move to optimal two-heaps solution (max-heap + min-heap)
  • Clean running code tested on various cases
  • Follow-ups on real-life scenarios (percentiles, bounded ranges)

Round 3 — Coding 2 / LLD (In-Person) — Hire

Problem: Design a Premium Cab Hailing Service for Uber

Difficulty: Hard

Focus areas:

  • Cab allocation logic (nearest available driver)
  • Clean, fully running code
  • Multithreading scenarios (concurrent ride requests, no double-booking)

Key learning: Master ONE language for LLD. The interviewer let me check syntax but it caused context switches. Fluency matters here.

Round 4 — Hiring Manager (Virtual) — Strong Hire

Difficulty: Medium

Behavioral + project-based:

  • Collaboration examples
  • Complexity of projects I've led
  • Working under tight deadlines and pressure

My approach: I structured my project explanations to naturally cover most behavioral themes (ownership, conflict, impact). Used STAR framework with concrete examples throughout. This round gives you a lot of control over the direction — use it.

Round 5 — HLD / System Design (Virtual) — Hire

Problem: Stock Price Change Alert System

Difficulty: Very Hard

Covered:

  • Push vs Pull mechanism (core discussion)
  • DB design and alert indexing
  • API design
  • Deep dive on trade-offs
  • Handling high-volume data (Uber loves this — Spark/Flink came up)

Overall Experience

One of the most fun AND grilling experiences. The in-person format for the first 3 rounds made it distinctly different from typical virtual loops.

Key Takeaways

  1. DSA: Expect variations of standard LeetCode patterns. Graphs (DFS, BFS, DSU) are Uber's favorite. The gap between Hire and Strong Hire is mostly about how well you handle follow-ups.
  2. LLD: Focus on a fully running solution. Master one language — syntax checking mid-round causes costly context switches.
  3. HM: You control the direction. Structure your project stories to naturally cover behavioral questions. Always STAR + concrete examples.
  4. HLD: Toughest round, but Uber asks from a relatively common set of questions. Focus on trade-offs and reasoning — rarely a single right answer. Prepare high-volume data topics: Spark, Flink, stream processing.

Tips

  • Get a referral — it speeds up recruiter outreach significantly
  • Practice graph problems heavily (DFS, BFS, DSU, Dijkstra variants)
  • For LLD, drill one language until syntax is muscle memory
  • For HLD, study Uber-specific patterns: real-time systems, high-throughput streaming, geo-based matching

Join 500+ engineers

Weekly interview questions + comp data insights. No spam. Unsubscribe anytime.

Powered by Substack. Your email is safe.