Interview for Data Engineering and Machine Learning Profile (20 Sep 2023) - For the position of Infosys Digital Specialist

Section 1: Programming

1. How much would you rate yourself out of 1 to 5 in these three:

Data engineering, ML Ops, Cloud

2. Broad concepts around Data Engineering and MLOps.

3. Write code to find the number of factors of a number.

import math
n = int(input("The number:"))

sqrt_n = math.ceil(math.sqrt(n))

l = set({})

for i in range(1, sqrt_n + 1):
    if n % i == 0:
        q, r = divmod(n, i)
        l.add(i)
        l.add(q)

print(l)
print(len(l))


Sample output:
The number:12
{1, 2, 3, 4, 6, 12}
6

The number:100
{1, 2, 4, 100, 5, 10, 50, 20, 25}
9

4. Tell what is the complexity of this code.

5. Can you suggest any optimization in it?

6. Write code to tell if a number is a happy number.

A happy number is a number defined by the following process:

- Starting with any positive integer, replace the number by the sum of the squares of it’s digits

- Repeat the process until the number equals 1 (where it will stay), or it “loops endlessly in a cycle” which does not include 1

- Those numbers for which this process “ends in 1” are happy.

Return true if n is a happy number, and false if not.

For ex: 19 is a happy number. Produces following sum of the squares of it’s digits: 19, 82, 68, 100, 1

And 2 is an unhappy number.

7. How would you identify an unhappy number for example: 2

A number is either a happy number or unhappy number.

We can create a list of all the happy numbers till 1000 and a list of unhappy numbers. Then preemptively stop on encountering one of those.

This way memoization would allow for optimization.

Section 2: Machine Learning

8. Which ML algorithm you are most comfortable with?

9. Can you take up questions on SVM?

10. The Machine Learning problem:

Let’s say you work in a financial institution, and you are given the task of using Support Vector Machines (SVM) to build a trading strategy for equities based on multiple features, such as moving average, volatility, and market sentiment.

Problem Statement:

To create an optimized SVM model that can effectively classify equities into “Buy”, “Hold”, and “Sell” categories based on historical and real time data.

Build an initial SVM model with a radial basis function (RBF) or polynomial kernel. Experiment with different parameters like the regularization constant (C), kernel coefficient (y), and others.

Discuss how are you going to do this.

11. How would you tune the hyper parameters of the model?

12. How would you use SVM on real time data?

13. What would be your strategy for feature selection?

14. What is RBF - Radial basis function?

15. What is matrix factorization with respect to SVMs?

Section 3: Cloud

16. Which cloud platform have you have used?

17. Which features of GCP have you used?

18. Which features of AWS you have used?

19. What is Elastic Cloud Compute or EC2?

20. What are steps of creating a project in GCP to use Buckets?

21. What are the steps of creating a project in AWS to use Lambda functions?

Tags: Machine Learning,Interview Preparation,Technology,