Qualcomm Interview Experience for ML and System Engineer

In October 2023, Qualcomm India visited our college for full-time recruitment across various job roles, encompassing Machine Learning, Multi-media, Computer Vision, System Engineering, Hardware, QP Digital-Analog, and Communication, among others. I will be sharing my interview encounter specifically for the Machine Learning position, noting that the experience is largely applicable to the Multi-media and Computer Vision roles as well.

1. Candidate Shortlisting :

The recruitment process was initiated with an online test, conducted by Qualcomm. This test consisted of three sections: an Aptitude section, a Basic C/C++ programming section, and an optional subject section, which encompassed topics like Machine Learning (ML), Signal Processing (SP), Computer Science (CS), and others. It’s worth noting that Qualcomm also considered the content of applicants’ CVs prior to shortlisting candidates for the online test.

The specific team in question was focused on hiring individuals for roles involving the deployment of Machine Learning inference on Snapdragon chips within smartphones. Additionally, the team was interested in candidates who could contribute to the optimization of models for efficient execution on Qualcomm’s Snapdragon chips.

2. Technical Round – 1

This stage of the interview was designed to assess your fundamental understanding of Programming, Machine Learning, and its theoretical underpinnings. The questions posed were quite elementary and didn’t delve deeply into intricate concepts. Alongside these, there was a straightforward C programming question included in this round. The points I’ve presented here mirror the way they were presented during the interview.

• Tell me about yourself in a 2-3 minute timeframe.
• 2 C coding question regarding trees and graphs.
• They go through my resume and asked me to explain one project in detail (technicals, scope, impact, etc).
• Asked about my internship, and what I did there.
• asked the Back-propagation algorithm in Dense Network, Back-propagation in Convolution Network, then asked me to write a C program for Back-propagation in Conv Network (I made an assumption that the gradient of output feature map is given at any layer). Asked me to explain this assumption. I drew a diagram to explain the back-propagation of CNN.
• Do you have experience in computer vision? Do you know the architectures of Inception, VGG?
• Explain the architecture of ResNet? Advantages of ResNet?
• How 1×1 Convolution reduces computation?
• Any questions? I asked about what work I will get and the typical workday.
• Explained about their team.

3. Technical Round – 2

• Write a program to and whether a number is a power of two. I gave different approaches but he wanted a “complement-based” answer which I was not able to deliver.
• How good are you with C and/or C++?
• Do you have experience with OS and Kernels?
• Explain Semaphore, mutex, locking, and threading.
• What is a deadlock in OS? What are the necessary conditions for a deadlock?
• Next, he asked me to choose one of C or C++ (I chose C and told him that it has been some time since I used C regularly).
• Write a program to and the intersection of two linked lists (in C). I gave a Hashmap-based answer. He asked me to code it. I asked him if I can assume a Hashmap implementation and he said yes. Coded it, and checked edge cases, syntactical errors, and logical errors.
• Write a clean implementation of the stack (any method). I implemented it with linked lists.
• How is memory allocated in C during malloc and calloc?
• How do you know whether a memory is in use or free?
• What are dangling pointers?
• Next, he wrote a code snippet and asked me to and out the error in the code, if any, and the
• type of error.
• What is a segmentation fault? He gave a few snippets and asked me if it would cause a
• segmentation fault.
• Any questions? I asked him more about the role and the kind of work they are expecting
• us to do. I told him frankly that I was not prepared for this kind of round.
• General guidelines when writing code during interviews:
• Make sure to use mnemonic names for variables and functions.
• Modularize the code.
• Use comments if necessary.
• Take special care of edge cases.
• Formulate and explain your approach. Discuss and agree on an approach before coding.
• Listen to the interviewer for hints and mistakes.
• After coding, walkthrough the code by speaking out to the interviewer.
• Explain the edge cases and explain why you chose to do what you chose to do

4. HR Round :

This round was just a quick phone call of ten minutes. The interviewer asked whether I was

explained the role properly during the previous interviews, my preferred location, and anything I want to know. She said she will inform the Placement cell about the result.

The interviewer evaluated my proficiency in C/C++ programming, OS concepts, ML concepts, and coding/problem-solving abilities. The initial round lasted around 50-55 minutes, while the second round extended for approximately 1 hour.

Verdict: Selected (Machine Learning Engineer)