Grab Interview Process 2026 — Rounds, Tips & Questions

You need to run a DAG that processes 10,000 files daily — one task per file. How do you design this in Airflow without creating 10,000 static tasks?

Design the driver-passenger matching system for a ride-hailing app operating across Southeast Asia, where network connectivity is inconsistent in parts of the region. How does that constraint change your design versus a similar system in a market with reliable connectivity?

Given an array of ride fare amounts collected over a day, find the maximum sum of any contiguous subarray of exactly k rides (a sliding window problem). Walk through the O(n) approach.

Write a query to find, for each city, the percentage of rides that were cancelled within 2 minutes of being booked, given a Rides table with city, booked_at, and cancelled_at.

Design a dynamic (surge) pricing system that adjusts fares in near real-time based on local supply and demand, without causing prices to flicker wildly for a rider mid-booking.

Your Airflow DAG needs to process each day's ride data only after all regional data files for that day have landed, but files can arrive up to 6 hours late from some regions. How would you design the trigger logic?

Two microservices in your architecture (Trip Service and Payment Service) need to communicate when a trip ends. Would you use synchronous REST calls or an async event, and why?

Grab operates across many Southeast Asian markets with different regulatory and infrastructure realities. Tell me about a time you had to adapt a solution because of constraints specific to one market or region.

How does Redis's GEO command family (GEOADD, GEORADIUS) let you find nearby drivers efficiently, and what data structure underlies it?

How would you test that a "cancel ride" API endpoint correctly refunds a rider only when the cancellation happens within the free-cancellation window, and charges a fee otherwise?