Round 1 – DSA and SQL (Technical)

General Introduction:

🔹 Discussed my previous tech stack.

🔹 Questions around data volume I’ve worked with, nature of my work, and the impact it had.

Coding Problems:

🔹 Longest Substring with Unique Characters

Input: str = "aabcdeeefijklmno"

Expected Output: "fijklmno"

Approach: Sliding window technique with a hash set to track seen characters.

🔹 Check if Two Strings are Anagrams (O(n), no sorting allowed)

Input: s1 = "tan", s2 = "ant"

Expected Output: Yes

Approach: Use a hash map (dictionary) to count character frequencies.

SQL Question:

🔹 Problem: From a product pricing table, find product names where the prices are strictly increasing over months.

Schema:

Fields: product_name, product_id, price, price_change_month

Sample Data:

Expected Output: b

Approach: Use window functions (LAG, ROW_NUMBER) or self-joins to compare consecutive rows and check for strictly increasing price trends.

Round 2 – Data Modeling and Spark

Project Discussion:

🔹 Detailed discussion on my previous project.

Spark-related questions focused on:

🔹 Data skewness handling

🔹 Code optimizations and partitioning strategies

Data Modeling Task:

🔹 Design a ride-booking app like Uber/Ola.

Created a Galaxy schema with fact and dimension tables to represent:

Users

Drivers

Rides

Payments

Locations

Ratings

OOP Code Design for Ride Booking System:

🔹 A simplified structure using classes and inheritance:

class Users(driver_available, fare_calc):
    def __init__(self, d_id, dname, location):
        self.d_id = d_id
        self.d_name = dname
        self.location = location
        self.fare = fare_calc.fare_func()


    def request_ride(self, location, vehicle_type):
        return driver_available.check_driver(location, vehicle_type)


class Driver:
    def driver_curr_location(self, driver_id):
        return (lat, long)


    def driver_veh_type(self, d_id):
        return v_type


class DriverAvailable(Driver):
    def check_driver(self, location):
        return is_there


class FareCalc:
    def fare_func(self, location, weather, vehicle_type):
        return calculated_fare


class Status:
    def driver_accept(self):
        return True/False


    def user_accept(self):
        return True/False


class Payments(FareCalc, Status, Users, DriverAvailable, Driver):
    def payment(self):
        if self.driver_accept() and self.user_accept():
            amount = self.fare_func(location, weather, vehicle_type)
            return amount, self.driver_curr_location(driver_id)
        else:
            return "ride_cancelled"

Round 3 – Situational Design + Hiring Manager (With Director of Engineering)

Discussion Highlights:

🔹 Introduction and detailed discussion around my role at Nielsen.

🔹 Responsibilities, impact areas, and future expectations.

Situational Design Problem:

🔹 How can Amazon Prime detect that a user logging in from different accounts and locations (e.g., India and US) is the same person, to show consistent movie recommendations?

Solution Approach:

🔹 Use probabilistic matching based on a combination of soft and hard identifiers.

Attributes Considered:

First & Last Name

Date of Birth

City of Birth

Phone Number (if same → high confidence match)

Email Address (if same → high confidence match)

Time of Birth (accurate but not user-friendly)

Personal security questions (e.g., favourite animal, father’s name)

Matching Flow:

Step 1: Match on phone or email → strong indicator (1:1 match)

Step 2: If not matched, apply layered filters on remaining attributes to narrow down potential matches.

Step 3: Prompt user confirmation via a non-sensitive UI message showing:

Last login device

Last login time

Location

Final Conversation:

Discussed expectations from the role, growth plans, and received the final offer 🎉