SRE / DevOps Engineer Interview Experience at Flipkart – 2025

Three DSA problems within a 90-minute window, matching the real reported structure of Flipkart's initial online assessment for infrastructure-facing roles. The first is a straightforward array manipulation problem (finding the maximum sum subarray of a fixed size). The second is a graph problem requiring BFS to find the shortest path in an unweighted grid with obstacles. The third, and hardest, is a dynamic programming problem involving partitioning an array into k contiguous segments to minimize the maximum segment sum - a working but not optimal solution within the time limit, with the complexity tradeoff noted explicitly, is a reasonable outcome.

A 30-minute technical interview conducted on Flipkart's HirePro platform. It opens with one DSA problem - reversing a linked list in groups of k - best solved iteratively while narrating the pointer manipulation out loud. The interviewer then pivots to a troubleshooting-style question: given a service that intermittently returns 502 errors under load but works fine at low traffic, what would the first three diagnostic steps be. A strong answer checks upstream connection pool exhaustion first, then load balancer health check configuration, then whether the service has a memory leak causing periodic restarts under sustained load.

The most substantial technical round of the loop, an hour long. It opens with the interviewer asking candidates to justify a specific deployment pipeline decision from their resume - why a blue-green deployment strategy was chosen over a rolling update for a particular service, and what the actual tradeoff was in terms of infrastructure cost versus rollback speed. If Redis appears on the resume, the conversation moves into Redis clustering: how Redis Cluster shards keys across nodes using hash slots, what happens during a failover when a master node becomes unreachable, and the consistency implications of Redis's asynchronous replication (specifically, that a failover can lose recently written data that had not yet replicated to the promoted replica). The round closes with a real incident the candidate handled: a specific downtime event, how it was first detected, the immediate mitigation applied, and what changed afterward to prevent recurrence.

A relatively relaxed closing conversation with the hiring manager, often finishing ahead of the scheduled time. It covers the technical ecosystem of the team, the specific line of business the candidate would be supporting, general team structure and reporting lines, and a light discussion of culture fit rather than any further technical probing.