Celebal Technologies: Data Scientist Interview Experience — OA + Interview(On-campus)

Data Scientist Interview Experience - OA + Interview

Hey everyone, sharing my recent experience with a Data Scientist hiring process. Posting this mainly to help people preparing for similar roles.

Round 1 - Resume Shortlisting

The process started with resume shortlisting. After getting shortlisted, I received the OA.

Round 2 - Online Assessment

Duration: ~1 hour

The OA had 4 main parts:

1. DSA - 2 Questions

There were 2 DSA questions:

• 1 array-based question
• 1 string-based question

Both were around Medium difficulty. Nothing extremely tricky, but you need to be comfortable with basic problem-solving and implementation.

2. Data Science / ML

This section was focused mainly on Classical ML.

Questions covered things like:

• Linear Regression
• Logistic Regression
• Classification/Regression
• Choosing the right ML algorithm
• Basic ML concepts
• General understanding of different techniques

There were also some basic Pandas and NumPy questions.

3. Python

This was mostly output-based.

Small Python code snippets were given and we had to determine the output.

The questions were mainly around Python fundamentals and syntax, so knowing how Python actually evaluates code is important.

Overall OA

The OA was fairly balanced between DSA + Python + Classical ML.

---

Round 3 - Technical Interview

Duration: ~35-40 minutes

The interview started with a quick introduction and discussion about my projects.

Around 20% of the interview was related to my resume/projects.

The remaining ~80% was Classical ML.

The interviewer covered a pretty broad range of topics, including:

• Linear Regression
• Assumptions of Linear Regression
• Logistic Regression
• Gradient Descent
• Bagging
• Boosting
• Different boosting techniques
• When to use different ML algorithms
• General ML fundamentals

The ML discussion was much more important than the resume/project discussion.

So if you're preparing for this role, don't just revise definitions. Be ready to explain the intuition behind algorithms and why/when you would use them.

---

Final Part - DSA

Towards the end, I was given a DSA problem.

The problem was essentially:

Find the second largest distinct element in an array containing duplicate elements, without using built-in functions such as sort().

Example:

[10, 5, 20, 8, 20, 15, 7]

Answer: 15

I had to both write the solution and explain my approach.

The expected approach can be done in a single traversal by maintaining:

largest
secondLargest

Complexity:

Time:  O(n)
Space: O(1)

The interviewer was interested in the reasoning as well, not just the final code.

---

My Takeaways

The process was pretty fundamentals-heavy, especially on the ML side.

For the OA

I'd prepare:

• Medium-level Arrays & Strings
• Python output questions
• Pandas + NumPy basics
• Classical ML fundamentals

For the Interview

I'd focus heavily on:

• Linear Regression + assumptions
• Logistic Regression
• Gradient Descent
• Bagging vs Boosting
• Random Forest
• Boosting algorithms
• Model selection
• Core ML intuition

Also, know your resume well. Even though the project discussion was relatively short, you should be able to explain what you built, your contribution, and the technical decisions you made.

Overall, I'd rate the process as moderate difficulty, with the interview being much more ML-focused than DSA-focused.

Waiting for the interview result :) (RESULT: NOT SELECTED) Hope this helps anyone preparing! All the best :)