FORESIGHT(2025) | Internship at NoBroker | Varun Agrawal

1) Brief introduction and description of offered role.

Hi, I’m Varun Agrawal, a Dual Degree student in Chemical and Financial Engineering at IIT Kharagpur. I’ve secured a Data Science Internship at ConvoZen AI, the conversational intelligence wing of NoBroker Technologies, for Summer 2025. ConvoZen leverages GenAI-powered solutions to analyze over 17 million conversations weekly, extracting actionable insights across domains like Sales, Customer Experience, and Compliance.

I believe my work will mainly focus on the NLP side — things like fine-tuning LLMs, experimenting with transformer architectures, evaluating models for our use cases.

2) How did you get into NoBroker/ConvoZen? What was the selection procedure?

I got selected through CDC. The process involved 2 technical rounds:

a) Round 1 (45 mins):

They grilled me on my CV. Some questions included:

• How TimeVQVAE works and the difference between AEs, VAEs, and VQVAEs.
• Limitations of GANs, encoding techniques.
• Deep dive into my projects and research work.

The round was purely ML-focused, no DSA (lucky for me).

b) Round 2 (45 mins):

This round focused more on the role-specific requirements:

• Heavy emphasis on NLP, especially the math behind Transformers, self-attention, embeddings.
• Questions from classical ML, bias-variance, model evaluation, etc.

Overall, both rounds tested depth in ML & NLP, not programming puzzles.

3) How to prepare for them?

• Start with Andrew Ng’s ML/DL Coursera courses if you’re a beginner.
• Focus on building research-level projects — these show depth and curiosity.
• For classical ML, know the concepts behind models, when to use what, and why.
• Read research papers if aiming for NLP-heavy roles.
• Be really strong with your CV — they deep dive into each line.
• For companies that do ask for DSA (like Trexquant), solve Striver’s DSA sheet.

Only a few companies like NoBroker skip DSA. So unless you know exactly where you’re applying, keep your DSA strong too.

4) When did you start preparing and what’s the ideal timeline?

I started my journey in March 2023 with my first DS project at IISc.

Get Student Welfare Group, IIT Kharagpur’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

My advice:

• Start DSA early, ideally in 1st or 2nd year.
• For DS roles, begin exploring projects and research at least 6–7 months before CDC.
• The earlier you begin with ML/NLP, the deeper your profile can be by the time companies arrive.

5) What difficulties did you face and how did you overcome them?

• Overwhelming scope: DS/NLP is vast and always evolving. New models like Transformers (2017) are still challenging even today. I handled this by doing impactful projects, and being solid in what I did.
• Staying consistent with DSA: Tricky when you focus on ML only. You either juggle both or hope for luck with ML-only companies.
• Very few DS internships on-campus. Be mentally prepared and explore off-campus options, hackathons, ML challenges

6) Are there any specific requirements like CGPA or department? Do PORs help?

• CGPA definitely matters. I had 8.59, but companies like Trexquant had 8.5 as a hard cutoff. A 9+ CGPA opens many doors, so maintain it if you can.
• Department can’t be changed now ;) but DS is open to all with the right skills.
• PORs don’t matter in tech-heavy roles. Instead focus on research projects, papers, hackathons, and public work like GitHub or Kaggle.

7) Who should ideally apply for this job?

Anyone targeting Data Science/NLP roles, especially those who:

• Are curious about voice AI, LLMs, or applied GenAI.
• Have done strong ML/NLP projects.
• Want to work on real-world conversational intelligence problems at scale. It’s perfect if you love both research and practical application.

8) What are some major points to include in the CV for this profile?

• Highlight NLP, ML, DL keywords.
• Showcase projects with metrics — e.g., accuracy, FID, MAPE.
• Mention research internships, papers, competitions.
• Keep formatting clean and readable. Projects like my TimeVQVAE + MaskGIT and GAN in Stock price were major highlights for me.

9) Final advice for students targeting CDC internships?

• DSA is a must for most tech companies. Start early.
• Prepare for SDE roles, even if your dream is DS — fallback options are important.
• Do not rely solely on CDC. There are great off-campus opportunities too, via:

i) Flipkart GRID

ii) Amazon ML Challenge

• Even if you don’t get selected in CDC, don’t lose hope — keep learning and applying. One good project can change everything.
• And also — I personally know friends who got great internships just by networking there. Reach out to people, ask for guidance, and apply smartly. If you ever need contacts or help, feel free to message me — I’ll be happy to connect you with someone or help however I can!