Yes, Sir, you can proceed. My interview.

Please start by introducing your professional journey in academia.

OK. Good evening, Sir. I'm Doctor A Stephens Hiaraj. I'm currently working as Assistant Professor, uh at Banariam and Institute of Technology, he wrote in Tamil Nadu. So I have uh. 10 years of experience in teaching and research so I have my area of expertise. It includes artificial intelligence and. That is mainly machine learning, deep learning, Transformers and also I'm currently working in quantum machine learning techniques. So. And I'm able to, uh, guide the students. And I have, uh, guide to the students, uh, whoever win the many competitions. And also I'm included, uh. My research in industrial consultancy work. And also. I gone with an research research expertise in FPGA implementation in. CNN implementation in FPGA. And also I'm currently working in the quantum computing and also the quantum machine intelligence also so. Uh motive, uh, is to design an uh. Uh, automatic. Uh and also in uh. Excellency. Uh, technique in agriculture and also in, uh, healthcare techniques. So that is my uh. Uh, intro so.

You've worked across machine learning, deep learning, and now quantum computing. Of all the foundational concepts you teach, which one do you most enjoy introducing to new students, and why?

So actually I have a a learn machine learning first. So the machine learning, uh, is an intelligent, I mean, uh. It's an interesting topic so because it includes all the statistics, techniques and also that. Few parts of data science and also in it includes a major part of training, testing and also an implementation in many real-time projects. So that is in basic work, so in which in, however, it can be evolved in the deep learning and also in quantum. So first of all I have I will be teaching machine learning to the students. So which which also favors to me because I have learned it first and also I have implemented in my PhD work and. In my thesis also. Uh, so that is, that is my interesting part and also because it evolves. Uh, it evolves from all the. Advanced learning concepts, it's been evolved from the machine learning, so that is why I have been liking more more. And also. It able to decides. The, uh, regressions. The classifications. And also in major part of an applications that includes from the machine learning technique. So I prefer a machine learning best for teaching to the students. So that is my wish.

You clearly have a passion for teaching machine learning, especially since it underpins so many advanced topics. When you first introduce regression versus classification to your students, how do you help them grasp the difference with a concrete example?

So can I proceed with next?

You clearly have a passion for teaching machine learning, especially since it underpins so many advanced topics. When you first introduce regression versus classification to your students, how do you help them truly grasp the difference between the two?

Yes, Sir, so. Uh, I will be uh, teaching, uh, first machine learning, uh, why we are going for a machine learning first so, and also I will be first telling the difference between the uh. Classification. And also from the regression so. Uh, I will give an real time examples of in classification like that. So how? We classifies the. Uh, fruits? According to the rotten and also in good fruits. And also how we can choose the classification part from the? Uh, that is an diseases, that is the brain tumor. Uh uh, within the normal. Uh, MRI images so and also, uh, for the regression I will be going for in real time example. That is for a prediction mainly. And so that is in climate predictions. So however with the existing temperatures. Uh, with existing, uh, wind speed. So how we are going to predict the climate, Uh, so, and also in uh. How we? Uh, how it will be accommodate? That, uh. The regression point of view. Uh, so and also when, uh. I will be teaching the regression from the linear regression so that linearly how will be adapted to the both the things so I can go for an linear regression so which? Will be supporting for the regressions also. And also I can explain the classification from that terms also from the linear regression the same. So that is been that can be a maybe an logistic regression. So I will be giving the. Clearview of differentiating the regression and also in classification in this manner. Uh, so such that the regression it can divide the. Uh, such that the classification can divide the. Whole part into two or major multi classification techniques and also the regression in the point of so how the X axis had been alonged within or cope up with the Y axis so such that the center line it can be designed by using the linear regression. So such a thing I can be working with the two techniques are these in classification and also from the regressions?

Professor, I noticed your 2025 paper in Scientific Reports—“Combination of gray level features with deep transfer learning for copra classification using machine learning and neural networks.” How has your experience from that research project influenced the way you teach classification tasks to your undergraduate students?

We have a simple binary classification. And also with an multi class classifications. So I will be explaining first that. Uh, so and for example, so in my research that is some combination of, uh. Uh, Gray level under. The deep learning features in the classification of Cobra. So what you have been acquitted. Coated so. Sorry, not acquitted. So that is an quoted. So that is uh. What you quoted. Uh, so such a thing in that. I will be using the binary classification. So first explains what is the need of an binary classification so such that students can able to understand. So what is the need of in classification so so we are going to proceed in the research. So in many terms so that may be the cobras and right co-path edition simply. Uh, so in that they. They have been added the. Uh, chemical so such that my objective is to, uh, differentiate the chemicalized the right copra. Dried coconut and also a non chemicalized, uh, dried coconut. So in that way. So I need to go for an uh. Faster ethics for classifying the. Thing So what is the need for the classification? So and I and I will proceed with how to classify it. So how to classify it? It is based upon the features. So So what are all the features we need to classify? The two things so that we need to think. To get the thought process from the students so. I I could imagine. I will give the assessment such that So what are the feature it can be obtained to classify the two things. So I will give in real time example so such that. It may be a cobra, It may be an banana leaf diseases. Uh, it may be another, uh, leaf diseases. Or otherwise I can go for a medical image. So first I will go within. Binary images, uh, so such that it can be, uh, this or that. So that classification I will be teaching. So in that. Uh, what are the features that can be extracted so that it can divide into? Two and also the. Uh, what are the models? It can be preferred for that classification to? So such that, uh, based upon the features, so how we can choose the? Uh, classification so that I I should be teaching to the students. So under on the. And the part of you so I can go for an programming, so the programming is second. The first I need to understand the concept of how we can. Uh, classify the. I have two things. So such thing uh here I, I I have uh classified the. Uh, bright Cobra. So such that. Uh, it can be classifying the. Sulfur fumigator and non sulphur fumigator copra which has been acquitted from the the images has been getting from the directly from the industries. So first I need to go for an image collections and also the preprocessing techniques and also the feature extraction technique. Amazon have and also I I have to teach how. Uh, how to classify? Based upon the features. Uh, and how to choose the models for getting the excellent accuracy in the terms? So in this way I will be teaching the classification so. Uh, based upon this research idea, so and I I will proceed with a multi class classification and multi label classifications so in which the people can understand because the binary classification is not only a main idea so. Uh, which can students offer for the research? So I I need to go for another introduction. So it is multiclass and multi label classifications. So that is the best way of uh. Going for an next reset Sir.

Let's switch gears, Dr. VIT places a premium on interdisciplinary research and funded projects. Given your background in AI and quantum machine learning, where do you see the most promising, fundable research directions over the next three years? Which agencies or grants would you target for external funding?

The opportunity to implement a quantum computing. Uh, lab in the.

Let's turn to how you teach. Suppose you were assigned a large introductory AI course—say, 200 students—and you couldn't use slides or a traditional lecture format. How would you structure a session to actually keep students engaged and help them wrestle with the core concepts?

Emerging techniques which will be going on. Next day. Uh, three years.

Imagine you're teaching that large introductory AI course—no slides, no traditional lecture. How would you design an active learning session so students actually wrestle with the material, not just listen?

And and etc. So that is, that's it.

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?

OK, so. The outcome assessment data. It normally comes from the. Projects. So and also from the exams. So the exams is normally has been going on in every colleges so which cannot be. Uh, interrupt. Umm, it cannot be interrupted so. I can go up with so if the projects idea. So our in the internship which the people going on. Uh, so from that the outcome is not if if I if I found the. Uh, VIP doesn't have any outcome sources in. For example, I'm saying if I. Not the. Found if I not found the. Project outcome. Or intensive outcome from the students. Uh, so in such a way so I will be. Uh, fix in the way so such that the project it should be on real time applications so that meet. Uh, support you? To the farmers or. Any other thing so either in Bay. The real time applications so which will be adapted. Which will be. Given to the farmers or real time consumers. So in that way the project it may be or. Either ways, you can go for an research project with which the outcome. It may be an uh. From the thing. So in in other form of view, so in our. Uh, college we have been uh. Uh, go, uh, go, uh, go with the go with that. Training often introductory classes, so such a way not introductory classes is a basic. Things learning SO in CSE department. So I can go up within skill settings often, uh. Uh, C language. Uh, maybe in Java? So our Python. Or it may be an uh uh. Uh, it may be an web designing so, or AR, VR or machine learning, artificial intelligence, whatever it may be. So I can divide that. Uh, thing because the students are. Go go going with an advanced concepts, not throwing the basic concepts. So for that I can introduce the. Concept often skilling. So in such a way in for each. Thing I can go with the leveling. So first level, second level, third level in such a way the people clear the first level, second level, third level. So they can either attempt the. Uh, outcome of the basic learning. So instead of going for an assignment. Uh, so Satya V the assignment they can. Copy from the chat CPT or whatever it may be. Uh, so instead. They can go within level so. If I have an C language this that is a basic of in computer science students, so I can divide that into N 5 levels. So in such a way I can go with an interaction of C. And functions so and also I can go within classes next structure, so whatever it may be so. With dividend of the class levels I can the students. I I made the students to attempt the exams, so that is in the real time. That is one year programming, so that is a real time. Scenario it will be given so in that the. The students have to work and they have to get the output. So that is the main thing. So instead of writing the assignment. The outcome part? Which cannot be reflected in good manner. So we can try. In the form of in skilling. Concept so success students can attempt the exams and also. Uh, not exams. So skill, uh, test. Uh, the they can attempt and also they can clear the. Uh, things. And also learn the basics. Without any assignments. And also without any. Uh, Lexus. That is a concept. And also with the. Uh, concept often. Self test. So that is the main way. Uh, we can. I can go with.

Picture this, Professor: a student files a formal complaint alleging your grading was biased, but at the same time, your department head is urging you to improve pass rates in your class. How would you navigate that situation?

Uh, and also we cannot face us the. Uh, examination point of view. So such a certain such a way I can convince my Hod. Uh, to retrieve. Process so such that the people can work with naturally and also we are giving some good things to the students such a way they will not go out and they will not. We biased to their employees or employer, whatever it may be, they can go. But then some. Uh, some sort of startup. So in that way, they will not be, they should not buy us. We should, uh, we should also teach. The students. Not technically, but also in morally. So that is my point.

Given your expertise in multimedia and artificial intelligence, do you have ongoing or past connections with industry that could help students access internships or placement opportunities?

OK. Yes Sir, I've been. Working with uh. An industry, uh, so in which, uh. Uh, his name is uh. And.

Let's shift to something a bit more practical. Can you walk me through a concrete example of an industry project or consultancy you've been involved in—what was your role, and how did your students benefit directly from that experience?

A medical expertise technique so such a way decision so LLM based clinical decision. So we have connected and. And this way. Uh, we had a meeting with the students who is working in the LLM with him. Under here Ajud. He asked that. Uh, the way. The students. Uh, presented to the clinical support he impressed in that way, so under. Uh, and he, they got the offer in the. Uh, some internship? Uh, regarding to the thing. And also, uh. And this is the one part of the LLM based clinical decision. So and also next thing we worked with an some hospitals regarding to the AI application. So that's such a way so far. Uh, for, uh, uh, monitoring the patients and also the patients in the canteen, how much time they are waiting, that is our objective. What they have given in the consultancy. Uh, so in this way, so the students have done the project. So and. Many. Uh, and many interviews the students have presented that. Consultancy what do the people and they got internship offers. Uh, so. And also the people working in the web development. So they have worked in the web development. I mean, my students have worked in the web development also and they have, uh, done with, uh, some, uh, international companies. Click on Select and Placements in the international companies. Uh, so this is the way. And also. Uh, so these are the projects has been done. Uh, during the consultancy and we got the opportunities of the internships and also the placements, so in another way. The students, my students have. Not the funding of two point 5,00,000 in May Solver UH. Thing. So in such a way. Uh, they got. Uh, first price about two point 5,00,000 and two point 5,00,000. Funding for the projects from the Gujarat Council of Institute and Technology from that is in DST India. Uh, so. And they got the project. And they have went and they have won the consolation prize. So in which they have, uh. Presented to the IIT Madras also in startup based companies. So in that they have presented and they have been got the internship regarding to that. Uh, projects. They have been uh. Two years. Uh, in that. IIT Madras working on the same projects that is in VLSI and also. The FPGA implementations.

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

I know Sir, I have been, uh. Here about VIP, so my friends are working there and my professors are mainly working there over there. Uh, so they've been excellent in the search. And that is the main point I'm going to appear on the interview. On 9th uh. Uh, April. Uh, in VIT. And so. I could I will be excellent in the research and also. And the the part of him consultancy projects and also the students development mainly so such that I can. Assured that I can train the students. Uh, that students can be uh, appeared in the many. I'm part of an international. And competition such that mainly over the SAH. Uh, other Technoxian? So, uh, and mainly. Uh, whatever it may be. The competition so the students can appear in that competitions and they can we know what the competitions so that is the main. More team, so the students uh guidance. For the projects. And also the research guidance for the searches.