Mad street Den visited our campus for 2 profiles, Software engineer and Machine Learning Engineer. I Applied for Machine Learning Engineer.75 students were selected for the online round.
Round 1: Online Round
It was conducted on the HackerEarth platform. There were 20 MCQ and 2 coding questions. Coding questions were of medium level. There was a negative marking in MCQ (+1, -1). MCQ were based on
- Statistics
- Probability
- Machine Learning
- DBMS
- Python
10 Students were selected for Technical interviews.
Round 2: Technical Interview 1
There were 2 panelists in an interview. They asked me to introduce myself and to explain any one of my projects mentioned in the resume. Then he did a lot of counter questions on my project.
- Is it always good to have accuracy as an evaluation metrics? What are other alternatives?
- What are precision and recall, when to use these as our evaluation metrics?
- How to use logistic regression for multiclass classification (one vs all )
- How to train a model where one image has more than 1 label.
- Advantages of YOLO Algorithm?
- Why do we use Anchor boxes in Yolo?
- What is cache memory?
- How will you replace the contents of cache memory if it’s fully occupied (Hint: LRU Cache)
- What is SQL?
- Types of joins in SQL?
4 students were shortlisted for the next round.
Round 3: Technical Interview 2
The interviewer asked me to introduce myself and then asked me to explain any project. He asked few project-related questions. After that, he asked me to share my screen and gave me a problem to solve.
- To find IOU ( Intersection over Union ) of 2 rectangles. The top left and bottom right coordinates of both rectangles were given as input. I was able to solve the problem but not in an optimized way.
- To remove duplicate entries maintaining the order of appearance. ( I solved it, but he was looking for a one-liner)
- Then interviewer shared a link, there were 3 SQL query-based MCQs and asked me to solve those questions.
Result: 2 Students were finally selected, and I was not among them.