I hold PhD degree in Computer science and engineering.

My research interest area.

Right, you mentioned your research interests. What foundational concept or core topic do you most enjoy teaching to undergraduates, and why does it resonate with you?

Uh, yes, uh, since I'm, uh, background is computer science and engineering and my. Visas interest area is natural language processing.

Okay, you’ve got a background in computer science and you’re interested in natural language processing. When you introduce NLP to students for the first time, what’s a core example or activity you use to help them grasp the basics?

And under gadget students.

Okay, you’ve got a background in computer science and you’re interested in natural language processing. When you introduce NLP to students for the first time, what’s a core example or activity you use to help them grasp the basics?

Yes, national language processing. What? What is the meaning of natural language? First of all, we need to understand the naturalism. Natural language means spoken language. So now it now in daily life. We have been using our. Our local languages, right? Now. For example, one application idea is. Machine translation.

Where one source language we need to. Translate into a particular target. Language. So it's a. But suppose you don't know the target language and you wish to. I know the what is the? What is the translation? In that target language, for example, I know English but I don't know Hindi. Right, so, so that so that is the thing. If you know that.

That's an interesting point. Could you elaborate a bit more on that?

Translation techniques so we can build up natural language processing applications so that people can use our applications so that they can give their input as an as an. Their input sentence. And they will get their output in their. Target sentence like Hindi, Hindi, Gujarati, Marathi any any. Language there one.

Now, focusing on your research in natural language processing, where do you see the most fundable directions over the next few years? Which grant agencies or industry partners would you target to support these efforts?

Right, so you’re talking about building NLP applications for translation between languages like English, Hindi, and Gujarati. If you were to seek external funding for this work, which agencies or industry partners do you think would be most interested in supporting your research, and why?

Yes, right.

Right, so if you're planning to develop machine translation tools, how would you go about building a sustainable research group around this at VIT—what kinds of collaborators, resources, or partnerships would you pursue to make it viable long-term?

OK, without losing the information. OK, so. In our target language, maybe?

Levi in Gujarati and Marathi. OK. So that is that is we are working on Indica languages. So that's all about my national level and international level. In language processing. Resource area.

Okay, so you’re working across a range of Indic and Arabic languages, and you mentioned using large language models for tasks like translation and summarization. Shifting gears to your classroom practice — say you’re teaching an undergraduate lab in NLP at VIT, but you want to move beyond traditional lectures and slides. What’s one hands-on activity or project you’d use to get a diverse group of students actively working with real NLP data?

Yes, I believe on project based learning. Mainly, uh, not like traditional. Lectures. Actually, I believe on hands-on skill. If they develop their hands on scale. For example. Any machine translation? How hot? Kind of tools, techniques available. We we should know OK what is corpus, what is? How to include? Data set OK How to build our own data set? OK, so that kind of. Techniques we can use Google collab.

Accreditation cycles can be demanding—if you joined our department and found that outcome assessment data was inconsistent across courses, how would you approach fixing that?

Uh, collect. Then how we can do the pre position part? OK, how can we use embedding techniques? OK then different model architecture will help. We need to implement like. Like a LSTM test 1 model. And also we can explode, uh, large language model, although we need. High end resource for large language model, but at least minimum level. Like Lama 3.1. Reply uh, because it's freely available. And so we can start. And we can. We can at least start and. So that they they they initiate the initial learning. With hands on practice. So I believe. Hands on practice and as well as. Learning project. By doing hands on project. They can learn quickly.

Professor, imagine this scenario: a student formally complains that your grading is biased, but at the same time your department head is pressuring you to increase pass rates. How would you handle this situation and balance academic fairness with institutional expectations?

Yeah, it's a very common. Question in every academic it it it is a. Challenging task. However, I will deal uh, nicely because uh. At the time. That is not like at the end. I will evaluate everything. It's like in continuous mode, right? If I. We like internal assessment. Assessment and then. Answer message man. So first of all, after taking some internal assessment, if I. Case that these students are not getting. Good marks. OK, then I then I can. I can separate them and I'm I will take. Extra class for them. So that there will there I will maintain the balance OK. And they will getting the uh. Better school. So the so that is 11 approach. Second approach is. Regularly like quiz and test I need to. You you I need to assess. So that. Before, before at attempting the mid SEM exam. There were they will prepare. What exactly we? We, we have. We are teaching. Or what exactly we are we need to focus? OK. So many, many, many. Things like for example any theoretical concept, if we know in practically, if we know some conceptually, for example gate level question. We need to. That we need to practice. For example normalization technique. So if I explain normalization techniques, different types of normalization, why you need? That normalization technique, for example. In order to normalize in order to. Uh, and normally. Any annually for for annually may minimize minimization and.

This uh. Redundancy issue. We will use the normalization technique. But they don't know.

Thanks, that's pretty detailed. What if your department head insists you raise grades across the board, even for students who haven't demonstrated improvement—how would you respond?

Dr., given your expertise in machine translation and NLP, do you have any industry connections or collaborations that could help translate classroom learning into internship or placement opportunities for VIT students?

Yes, I have been associated many. Startup organization like Leo Feliz Instrument. 651 Boss Pvt. Ltd. So also I'm I I have been as. Have a collaboration like. Barack and Bridges, Bangalore. So so that kind of. From collaboration, that kind of. Uh, the resources we can use. We can also invite those expired. Uh, at the at the same time, so, so that they may get. Internship opportunity. They are OK if you, if you, if, if you, if. They crack their interview. So before interview. We can. We can train them how? What exactly they they actually looking? Which kind of candidates? OK, so basically I can hire, I can invite. Those people. Who can also.

Oh help me for training that it. Project based training. So that they can. Achieve, uh, internship.

Let's talk about your experience with student evaluation and exam duties. Can you walk me through a time when you had to design or grade an exam for a large class—what steps did you take to ensure consistency and fairness?

Alright, we're wrapping up. Is there anything you'd like to ask about the role or the team?

Yes, yes. We can choose. Small projects. Specific problem statement. For example what? Or for for example if we choose any. Machine translation task. Although. There are beautifying public statements, are there? OK. So we need to do that one, uh. What you eat within the feasible timeline. And you can set. And they can start their work accordingly. So first of all. For any topic, not only machine tester. So it's some. Question answering. Information retrieval, any task. First of all, they have to do.

Literature. OK, if their literature review is is. Is very strong. Then they can highlight the resource gaps. And motivation then? Highlight. 3 objectives, one truth or. 2 objectives. Then do I achieve that objectives they can start their work? And that. Within the feasible timeline, they can. Surely complete, I believe.

Alright, we're wrapping up. Is there anything you'd like to ask about the role or the team?

Alright, we're wrapping up. Is there anything you'd like to ask about the role or the team?

Yes. So my basically my point of discussion is. I should involve in teaching or? Results because I have in. I have been, uh, interested in both, not only teaching. Research and teaching I I. I have experience and also. I need to explore. Uh, efficiently in bit. So that is my concern. Not only teaching. But also. Research.

That's a good point, Dr. The role here does involve both teaching and research—faculty are expected to contribute actively in both areas, with flexibility to shape your own balance over time. Is there anything else you'd like to ask?

Well, for now it's OK.