Recently, I had the chance to interview for a Data Science Internship at Tata Communications through on-campus hiring drive 2025. Though I wasn’t selected, the journey was insightful and taught me key lessons I’ll carry forward.

Let me walk you through the process.

Job Details
Role: Intern — Data Science
Duration: 5 months (till June-end)
Mode: Hybrid (3 days WFO)
Stipend: ₹35,000/month for B.Tech and ₹45,000/month for M.Tech
Location: Tier-1 cities like Mumbai, Bangalore, Delhi, Chennai
Eligibility: B.Tech/M.Tech — Data Science & AI (no active backlogs)
Tech Stack: Machine Learning, Deep Learning, EDA, LLMs, Computer Vision

Selection Process
Around 81 students were initially shortlisted for the coding assignment based on a Microsoft Form shared by campus. And 38 candidates made it to the virtual interview scheduled for a duration of ~30min.

Virtual Interview
The interviewer was friendly and started with a casual “How are you doing?” followed by my self-introduction.

He then gave me a choice: Classic ML, NLP, or Computer Vision — I chose Classic ML.

Questions Asked:

1. What is Linear Regression?
   Follow-ups:
   What is the equation?
   What do slope and intercept represent?
   How do you evaluate the model?

2. What is Gradient Descent and its types?

3. What is a learning rate?
   Follow-ups:
   Can you explain its equation?
   What should be the optimal value of learning rate?

4. When does a curve “converge” and what does converging mean?

5. What is an optimizer with examples and when it is used?

After an awkward pause, the interviewer transitioned to live problem solving and asked me to open my favorate IDE.

Problem: Write a function to calculate sqrt(x) without using inbuilt functions. I chose to code in Java.

I implemented a basic solution using a for loop, but the interviewer made me think wanting an optimized approach using Binary Search (to reduce time complexity to O(log n)).

Unfortunately, I hadn’t practiced problems related to Binary Search and couldn’t come up with the logic. To be honest, I didn’t even know that Binary Search can be applied to solve it at that time.

The interview ended with: “I think I’m done with you.”

Outcome
I didn’t get selected.
Honestly, I lacked clarity in ML fundamentals and didn’t prepare DSA patterns well enough. But this was my first ever Data Science interview — and it helped me identify my weak spots clearly.

Key Takeaways
✅ Master the basics and inner workings of ML Algorithms — Don’t just memorize definitions — deeply understand what’s happening behind the scenes: the math, intuition, and real-world use cases.

✅ Practice DSA to identify patterns — Consistent DSA practice helps you think in terms of patterns. Even if you can’t code it completely, explain your thought process to the interviewer — that shows maturity.

Final Thoughts
While I didn’t crack the internship, I’m genuinely grateful for the experience. It gave me a real glimpse into what’s expected and where I need to grow.

If you’re preparing for Data Science roles, I hope this gives you clarity and motivation to keep pushing forward.