Account Balances - Challenge Questions

General Instructions:

• There are 2-4 parts in total.
• The parts are meant to be completed sequentially.
• Optimization is not the main focus. The goal is to get your code running successfully and passing tests.
• Time is limited (around 30-40 minutes). Try to solve all parts within the given time. Get a working solution first before trying to optimize.

Part 1: Compute Final Balances

Problem

At a company like Stripe, we move a lot of money around. It is often important to know how much money an account has at a given time.

Implement a function that computes the account balances given a list of transactions.

Each transaction contains:

• account_id — The account identifier
• timestamp — The transaction timestamp
• currency — The currency code
• minor_unit — The transaction amount in minor units, such as cents for USD

Transactions are provided in strictly increasing order of timestamp.

Requirement: Return the final balance of each account, excluding accounts whose final balance is zero.

Example Input

account_id,timestamp,currency,minor_unit
acct_123,1,usd,800
acct_321,2,usd,100
acct_123,4,usd,-300
acct_321,5,usd,-300
acct_321,9,usd,-800
acct_321,10,usd,1000

Expected Output

acct_123: 500

Explanation

• acct_123 = 800 - 300 = 500
• acct_321 = 100 - 300 - 800 + 1000 = 0

acct_321 is excluded because its final balance is zero.

Part 2: Prevent Negative Balances

Problem

Modify the solution so that an individual merchant balance can never go below zero.

For every transaction:

1. If applying it would make the account balance negative, reject the transaction.
2. Otherwise, apply it normally.

Requirement: Return the rejected transactions and the final balances.

Example Input

acct_321,2,usd,100
acct_321,5,usd,-300
acct_321,9,usd,-800
acct_321,10,usd,1000

Processing Logic

• +100 -> 100 ✓ accepted
• -300 -> -200 ✗ rejected
• -800 -> -800 ✗ rejected
• +1000 -> 1100 ✓ accepted

Expected Output

Final balance:

acct_321: 1100

Function Template

// Complete the 'compute_account_balances' function below.

function compute_account_balances() {
    // Your code here
}

function main() {
    compute_account_balances();
}

Complexity Expectations

• Time: O(N)
• Space: O(A)

Where:

• N = number of transactions
• A = number of unique accounts

The interviewer gave me a GitHub repo and asked me to open GitHub Issues for the issues present in the repo. There were around 3–4 issues that needed to be identified and documented.

The interviewer asked me to clone the repo locally, open it in VS Code, and start working on it. I preferred using JS since I’m more comfortable with it; I don’t have much experience with other languages.

This round was honestly quite hard, especially if you don’t have hands-on experience with GitHub and JavaScript-level coding. The expectations were kind of production-level.

I still had the repo locally that I cloned during the interview, so I’ve attached it here. It might be helpful for you all while preparing.

JSON Parsing & Data Extraction - Challenge

General Instructions:

• You will be working with GPS data encoded as GeoJSON.
• Optimization is not the main focus. Get your code running successfully first.

Part 1: Parse JSON

Problem

Let's parse some JSON data.

(Note: If you're using a statically typed language such as Java, Kotlin, C#, C++, or Go, you may want to review how to handle dynamic/unstructured objects in JSON, though you may map it to structured objects if you prefer).

Your Task:

1. Write code to extract the coordinates from ride-simple.json.
2. Print the first ten coordinate pairs to stdout.

There are almost 500 points in the file, but we only need to see the first 10. This is GPS data of a bicycle trip to Stripe's Seattle office, encoded as GeoJSON.

Input

The input is provided in a local file named: ride-simple.json

The JSON contains a GeoJSON object with GPS coordinates. The coordinates are available under the following path: rideData.features[0].geometry.coordinates

Each coordinate is represented as an array containing two numbers: [longitude, latitude]

Expected Output

Print the first 10 coordinate pairs to stdout.

Example:

First 10 coordinates
[-122.2851, 47.55126]
[-122.28531, 47.55126]
[-122.28535, 47.55126]
[-122.28565, 47.55126]
[-122.28575, 47.55126]
[-122.28598, 47.55126]
[-122.2862, 47.55126]
[-122.28631, 47.55126]
[-122.28687, 47.55127]
[-122.28767, 47.55127]

Note: The exact coordinates should be dynamically extracted from ride-simple.json rather than hardcoded.

Function / Starter Code (Node.js Example)

'use strict';

const fs = require('fs');

try {
    const rawData = fs.readFileSync('ride-simple.json');
    const rideData = JSON.parse(rawData);

    const coordinates = rideData.features[0].geometry.coordinates;
    const firstTenCoordinates = coordinates.slice(0, 10);

    console.log('First 10 coordinates');

    for (const coord of firstTenCoordinates) {
        console.log(coord);
    }
} catch (error) {
    console.log('Error processing JSON file:', error.message);
}

Running the Code

If using Node.js:

node output.js

What Changed? (Subsequent Step Variation)

A later version of this task removes the .slice(0, 10) limitation.

Before (Part 1):

const firstTenCoordinates = coordinates.slice(0, 10);
for (const coord of firstTenCoordinates) {
    console.log(coord);
}

Updated Version (Part 2+):

for (const coord of coordinates) {
    console.log(coord);
}

The updated version loops directly over the coordinates array and prints all ~500 points.

Requirements Checklist

• Read and parse ride-simple.json.
• Extract the coordinates from the GeoJSON structure.
• Print only the first 10 coordinate pairs for Part 1.
• Do not hardcode the coordinates.
• Use any preferred JSON/data structure approach.Tips: I applied through a referral and gave the OA( having some system design level qs with 3 parts ). After two weeks, I got my R-1. Then, after another month, I got R-2 and R-3 on the same day. Now, I’m waiting for the final HR round.

In the last round, I fumbled quite a bit and wasn’t able to solve more than 1.5 parts. The interviewer’s expectation was around 3 parts.

So, hoping for the best! 🤞

Waiting for this round 😬