Summary
I applied for a Software Engineer - Full Stack role at Meltwater with 3.5 years of experience, securing interviews after an application via Naukri. After successfully navigating an HR screening and a technical screening (covering React and Data Structures & Algorithms), I was unfortunately rejected in the final Hiring Manager round due to difficulties with a follow-up question.
Full Experience
I applied for a Software Engineer - Full Stack position at Meltwater through Naukri, bringing 3.5 years of experience to the table. The interview process consisted of three main rounds: HR Screening, Technical Screening, and a Hiring Manager round.
My journey began with an HR Screening Round. We delved into my recent project experiences, the technical challenges I've encountered, my specific contributions within projects, and any live production issues I'd successfully resolved. A notable question also focused on the database technologies YouTube employs and the rationale behind their suitability for its specific use cases. I was pleased to be shortlisted for the next stage.
The Recruiter informed me that the subsequent Screening Round would be split into 20 minutes of React assessment and 30 minutes dedicated to Data Structures and Algorithms (DSA).
In the React section, the interviewer presented me with 5-6 code snippets sequentially, asking me to complete them. The core emphasis was on my understanding of React hooks and my ability to identify and apply hooks like useReducer, useCallback, and useMemo in various scenarios. While my hands-on experience with React was somewhat limited, I managed to answer approximately 50-60% of the questions.
Following the React assessment, the DSA section presented two problems. The first involved processing a peculiar dataset structured as strings like [robot1_tool1, robot2_tool2, robot1_tool3, ...], with the goal of outputting a mapping of each tool to the robots that utilize it (e.g., tool1 - [robot1, robot2]). The second DSA question required me to find all root-to-leaf paths within a directed graph, which I understood to involve identifying nodes with an in-degree of 0 and then performing a Breadth-First Search (BFS) or Depth-First Search (DFS). I successfully cleared this screening round.
Finally, I advanced to Round 1 with the Hiring Manager. This round involved two algorithmic questions. The first problem asked me to find the earliest 0-indexed position in an array of point differences from a football match where a team transitioned from a trailing or tied state to a leading state, returning -1 if no such transition occurred. For example, for [0, 1, 0, -1, 0, 1, 2, 1, 0], the output should be 5. The second question was a follow-up to the first, introducing an array of rectifications (time and score updates) that would affect all subsequent values in the point difference array from the given time onwards. After applying these corrections, I needed to solve the original problem again. I struggled to fully grasp the follow-up question initially, which seemed to disappoint the Hiring Manager. Regrettably, I was rejected after this round.
Interview Questions (5)
Explain what database technologies YouTube uses and justify why they are a good fit for its specific use cases.
Given a dataset of strings where each string represents a robot and a tool it uses (e.g., robot1_tool1), process this data to output a mapping of each tool to a list of robots that use it. For example, given [robot1_tool1, robot2_tool2, robot1_tool3, robot3_tool4, robot2_tool1, robot3_tool3], the output should be formatted as tool1 - [robot1, robot2].
Given a directed graph, identify all root nodes (nodes with an in-degree of 0) and then find all distinct paths from these root nodes to any leaf nodes. For example, given an adjacency list like A -> [B, E, G], C -> [D, F, I], H -> [F, M, N], you need to output all such paths. The problem suggests finding in-degree 0 nodes and performing a BFS or DFS from them.
Given an array of integers representing the point difference of a team in a football match (where positive means leading, negative means trailing, and zero means tied), find the earliest 0-indexed position (index) where the team transitioned from a trailing or tied state to a leading state. If no such transition occurs, return -1. For example, for input [0, 1, 0, -1, 0, 1, 2, 1, 0], the output is 5.
This is a follow-up to the previous problem. In addition to the array of point differences, you are now given an array of rectifications. Each rectification is an entry [time, score_change] indicating that at the given time (index), the point difference needs to be updated by score_change. This correction should affect all subsequent values in the point difference array from that time onwards. After applying all rectifications to the original point difference array, the task is to find the earliest index where the team transitioned from a trailing or tied state to a leading state, similar to the original Q1.