Freshworks | Interview (1st Round) | Bengaluru | Data Scientist

Interview Context

I appeared for a Data Scientist role at Freshworks in January 2025. I had 3.5 years of industry experience as a Data Scientist in a mid sized company. The first round was a technical round consisting of two coding questions and a set of ML + NLP theoretical questions.

Below are the questions I was asked:

🔹 Q1: Simulate a Two-Die Casino Game

Problem Statement: You are working for a casino and need to simulate a two dice game. Write a function that simulates the sum of rolling two six-sided dice, but you are only allowed to use one random function.

Constraints: • Use only one call to a uniform random function like random.randint(1, 6). • Return values should range from 2 to 12, with correct probability distribution.

Follow up: Avoid multiple calls to random function

🔹 Q2: Maximum Minimum Thickness of Slab

Problem Statement: You’re given an m x n matrix, where each cell holds the thickness of a slab. You need to move from the top-left cell (0, 0) to the bottom-right cell (m-1, n-1), moving only right or down.

The goal is to maximize the minimum thickness encountered along the path.

Input:
[
  [5, 1, 7],
  [4, 8, 5],
  [2, 3, 9]
]

Output: 4
Explanation: Path = [5 → 4 → 8 → 5 → 9], minimum along path = 4

🔹 Recommended Practice (as per HR):

• Number of Islands (Graph traversal / DFS-BFS problem): https://leetcode.com/problems/number-of-islands/description/

🔹 ML / NLP Theory Questions:

1. What is a reranker?
2. Difference between Bi-Encoder and Cross-Encoder?
3. How is BERT different from GPT?
4. What are activation functions used for?
5. What is regularization and why is it important?

Update:

Second Coding question was not very easy, I could discuss the approach only.