Razorpay | Senior Software Engineer | Interview Experience | April 2025 | Rejected

Exp : 2 Years, 10 Months
Current Comapany : SaaS Startup in Blr
Application Process: Applied via Referral

Round 1 - Machine Coding
Duration : 90 Minutes (Inlucding Intro, Question Explnation, Running the test cases and explaining the question)

Basic into and then interviewer gave this question

Expecation - Running Code

Your organization has started a new tech blog with interesting tech stories and you’re responsible for designing and
implementing an in-memory search engine, supporting the search functionality on the blog content.
Requirements:

1. It should be possible to create a dataset in the search engine.
2. It should be possible to insert and delete documents in a given dataset. Each document is simply a piece of text.
3. It should be possible to search through documents for a search pattern in a given dataset.
4. It should be possible to order the search results

This is a frequently asked question at Razorpay and has been shared by multiple people on leetcode already.
I presented a simple CRUD application. Sharing the code for Serach Service in the end of the post.
I was able to run all the 4 requirements. The intervievier then asked me about inverted index & if I know about it. My solution didn't involve creating inverted index, but I did know what it is and explained the apporach via inverted index in the end.

If you are prepping for Razorpay, I highly reccommend soliving this question via Inverted Index and have some knowledg about suffix & prefix tree.

public class SearchService {

private DocumentService documentService;
// pattern = search Query
public List<SearchResult> search(String blogName, String dataSet, String pattern, SortOrder sortOrder){
    Collection<Document> documentCollection = documentService.getAllDocuments(blogName,dataSet);
    List<SearchResult> results = new ArrayList<>();
    for(Document doc : documentCollection){
        int freq = countOccurance(doc.getText(),pattern);
        if(freq>0){
            results.add(new SearchResult(doc.getId(),freq));
        }
    }
    results.sort((a,b) -> sortOrder == SortOrder.ASC ?
            Integer.compare(a.getFreq(),b.getFreq()) :
            Integer.compare(b.getFreq(),a.getFreq()));
    return  results;
}
private int countOccurance(String text, String pattern) {

    int count = 0 , index = 0;
    while (true){
        int foundIndex = text.indexOf(pattern,index);
        if(foundIndex==-1){
            break; // no more to mach
        }
        count++;
        index = foundIndex + pattern.length();
    }
    System.out.println("Count of pattern " + pattern + " in content " + text + " = " + count);
    return count;
}

}

Verdict : Positive and next HLD round was scheduled.


Round 2 - HLD Round
Duration : 60 Minutes
Question - Design Leetcode - Foucs on LeaderBoard Design.
I presented a basic CRUD, some security and caching optimzations.
The interviewer had various questions towards the end but we were running out of time and I was not able to answer all.

This is not the best Design. I'm sharing whatever I was able to draw in the interview w/o optimizing.


There is a lot of stuff which was verbally explanined to the interviwer which is not written in this HLD diagram.






Some questions -
What if AWS Lamda goes down?
What if Redis Goes down?
AWS Lambda Cold Start?
How will we run test cases?
Stale Data?

What is the reason behind choosing a NoSQL/SQL DB? My response for this was that it really doesn't matter which DB we choose, because for our use case both SQL & NoSQL would work perfectly fine. This is a completely fine answer in my opinion but I'd strongly suggest here to have some points to explain some pros of the DB you are choosing.Just let the interviwer know you know about the DB you are choosing.

There is a lot of stuff which was verbally explanined to the interviwer which is not written in this HLD diagram.

Called the HR for Feedback - Rejected

Learning - Need to do basic CRUD design faster in HLD to be able to optimize the system and answer interviwers questions later.