Good day, Dr. V. Krishnakumar. I'm Arjun. I've had a look through your background and your professional record. I'm interested in hearing more about your work and how you see yourself fitting in here with our faculty. Ready to start?

Yes, Sir.

Please go ahead.

Yeah, umm, myself, I'm Doctor G Krishna Kumar. I have completed my BE Electronics and communication engineering in Doctor Jeep College of Engineering, uh, thirdly, which is Commander and university of Anna University Chennai. And after that I have joined in Kona College of Engineering and I have did my PG communication system in Jayaram College of Engineering, uh, under Anna University. And I have completed my PhD in January 2025 in the area.

Background and I have worked as a working association professor in Kangana College of Engineering which is located at Tiruchi since I'm there from the 2008.

For networking, these are all our. These are all my academic.

Thanks for sharing that, Professor.

Dr. Krishnakumar, let us begin with some technical aspects of this interview regarding your expertise in Artificial Intelligence, Machine Learning, and Data Science. Could you explain how supervised learning differs from unsupervised learning and which key factors you consider when deciding which method to apply to a specific problem?

Yeah, yes, actually supervised landing we need to fix some algorithms and unsupervised algorithm without knowledge of being the background of the problem, we can able to apply the algorithm over that and we can able to solve that issues. The main key factor is in supervised learning, we have to train the models and.

A basic knowledge about the database only we can able to solve that. But unsupervised learning we can able to train without training a model we can able to do that.

So we have to we have some set of database from using the.

Thank you for your response, Professor. Could you elaborate specifically on an application where you've used supervised learning, and explain why this method was ideal for that situation?

Yeah, uh, actually suppose the same thing, uh, the supervised planning, if I am uh, completed over that for uh, uh, like, uh, disaster management. Uh, so we know that the, what are all the key factors which were affected. For example, we required hospitality and we require emergency and some, some different cases are there. So we know that these are all the requirements by the concerned umm team. So big teams which was affected. The disaster. So we have to train the models and we have to fix the certain threshold over that and we can implement the supervised planning over that because we know the actual threshold and what is the actual requirement.

For those people, this were affected in the disaster area. So we can train by using some for example drone-based implementation, geographic landing. So over that we can able to implement the supervised planning.

Understood. That's an interesting use of supervised learning. Could you briefly contrast this with an example where unsupervised learning would be more suitable, perhaps in a similar domain or another field?

Yeah, yes, Actually, uh, the unsupervised learning means, uh, we don't know any exact background over that, but the same thing, for example, uh, even if I'm, if I'm analyzed that, whether in a disaster management we require some hospital hospitality means, uh, we can adopt or we can give some options to the doctor to look over there. The doctor can simply analyze that patient, for example, those who are affected using the disaster. There is no special required training over there because he already have the knowledge how to train the how to teach the sorry, how to give aid to that people. So I hope that I'm once we are establishing the supervised planning under that, we can bring the unsupervised landing over that with a proper guidance over that by simply allocating.

Suppose maybe, for example, I said that as a doctor.

Got it, Professor. Thank you for providing those examples. Let's move on to your teaching abilities. How do you approach explaining complex concepts, such as backpropagation in a neural network, to students who are new to the topic?

Something it's actually, I can just explain the overview about the back propagation model. Actually back propagation model, we have a set of neurons over there, neurons over. That means so we can take our body as one example and we can have the more number of neurons are there. So each and every part is connected with the neurons. So the brain is the main. Main like server so. So suppose if I want to push my hand from one side to another side or One Direction to another direction, what it implies is I can it was induced in my mind and it can transform through the neurons and I'm doing all those activities. So once if you are thinking over that such a problem, we have to stimulate over the neurons and. Which model is suitable for us or which hand which I want to?

Take or which which hand I want to move over that so, so, so this this type of backgrounds or this type of a simple examples I have given over there. And after that one step by step, I can enter into the back propagation model and I can explain the set of suitable postulates and basic theory and basic equations and how we are equating the LHS part for example, LHS spot and the RHS part like that. And that can complete my class.

Thank you, Professor. Your neuron analogy is an interesting way to make backpropagation relatable for students. Could you share how you evaluate whether students have fully grasped such a concept after your explanation? For example, do you use specific assessment strategies or exercises?

Ah, yes, actually the main objective is uh, what we are saying, the information could reach the students and the evaluation is the major process. Whether I need some satisfaction, whether I have delivered the information in a correct way, whether those information is reached to my students in a proper way. That is my main expectation of me as, as being a good teacher. And so I, I can use some on the classroom itself. I can. Heard something, some something, for example, some set of questions. Those questions can be in a different way, different layer. Like Mike. It might be like simple objective questions. Simple analytical questions are like future, future evaluation or future direction pushing in order to make the students to think. For example, I can take my Google form as my simple tool and I can share those set of.

Information in the class itself and ask the students to answer those questions. From through that on the spots itself, I can able to analyze whether the student can understood the my topic or not. Suppose if so some set of result may be generated. Those results I can go through each and every result whether which part the students are not able to understood. Some that if it is required if I have a time being during those alerted our means, I can able to explain once again whether some of the students are not.

Able to understood those topics and I I think that this is 1 better way and some more number of online platforms are there those platforms also I will use for example Meta. I think online, online assessment is also there as being a so I'm being some I'm preparing some setup questions and load over that and ask the students to answer. And after analyzing I can able to understand whether my students understood or not.

Thank you, Professor. Your approach of using immediate feedback and analysis through tools like Google Forms is a practical way to ensure understanding. Now, regarding your experience with guiding student research, could you share an example of a student project you mentored and your role in supporting their work?

Uh, yes, Sir, actually, Sir. Umm, uh, in the recent years, in the past, I have, uh, guided one student. I'm like, he actually, the farmer is requiring, uh, the farms to make a fertilizer to spread over that at the same time. And those fertilizers are using the fertilizers that will affect the human health also. So in order to prevent that, we are creating, uh, some.

Mom, the yay based drone. Yay based drone means we are we are. We have give some set of images for the forms and those forms require the fertilizer to be spreaded over those fertilizer can be fed along with the drone instead of instead of a human involvement, the human can simply sit in a particular place and as the drone to spread that over that. So that requires fertilizer. For example the farm consists of a different types of like yields are there. We may not able to use the same fertilizer over all the fertilizer because each and every fertilizers like is different for different yields. So first we can pictorially analyze all those images and we can apply some algorithm and categorize all those. Yields. And we can separate OK for particularly from this area to the that area, we can able to use the QAS technique geographic implementation over that. And once the drones analyzed okay, for this field or the this form, we want to give this type of fertilizer and we want to spill over that yield like that. So I, I have assisted the students to make this and.

Yeah, and I have completed this project also.

Yes, actually it's a good question. Thank you professor to make me to think over that. And 1st actually, I when I'm when I'm giving some information, whether this was my expectation that I'm saying to my students, they so actually my students are very first graduate student. First graduate student means one of the, I think one of my student father is a farmer. He's simply asking the questions to his father itself and he analyzed some data sets. OK, these are all.

Thank you, Professor. That's an excellent example, showcasing both your mentorship and the integration of technology to address real-world challenges. Could you expand on the specific challenges you faced while guiding this project, particularly in terms of aligning the student's technical approach with the practical needs of farmers?

Problems that I have facing like that and the same information we we have sit together along with the student and what are all the things we're from as a as a engineer or as a particular professor what, what are the ways to analyze the problems. First, we have categorized the problems and so one thing is for example, when we implement this project over a land area that is the possibility of some power restrictions over that and also.

We need to train the farmers over that because they don't know how to operate and all those things. For that we we are training the farmers, we are giving some, some amount of input on how to manage this position of the drone and everything remaining everything we we don't know. We actually once we have designed and we are, we are simply protocoling everything. Protocol everything means we are making the step by step process. So these are all the steps. We want to do and these are all the things. Which you could not do like that. We are simply saying some set of charts over there and the technical challenges what we are facing is the main thing is the power construction. Power consumption is suppose if area is somewhat more means on that case we may not able to do the make the drone to use over a long period of time. That is the major challenges we have faced that is mainly due to the battery restrictions. But we have worked over that in order to create some lightweight because the. Wants to carry the tank over for a certain.

Distance from the ground. So we want to peculiarly focus over that and we tried to implement over for a longer area by simply adding more number of drones over there. Thank you.

Thank you, Professor, for detailing the challenges and innovative solutions during the project. It demonstrates a proactive approach to both technical and practical concerns. Let us now evaluate your research impact. Could you discuss one of your published papers, particularly the research problem you addressed, and the methodologies you employed to derive conclusions?

Ah, yes, actually I hope I have completed the PhD in the year of January 2025 and so I can take the my research over there. Disaster management, actually disaster management may be a maybe a like it may be a man made or maybe a natural, natural disaster. And so I I can think over that how to. Transfer or how to implement the technology over the disaster management that is the.

Main objective of my research work and I have split my work into three parts. 1 is first part is in order to whenever we are implied to implementing the mobile ad hoc network over the disaster network, that is the possibility of implementing the management. Actually what the main drawback that I have analyzed is first one to initiate the communication is the major problem in the disaster area. Why? Because of the disaster acquires means.

Everyone is afraid to go there and how to initiate the process is the major problem. So I have taken that first we want to establish the communication over the disaster network. And the main thing I have found out is the heuristic approach is the best for me. And under the heuristic approach, I have taken a geographic drone based communication. Geographic drone based communication is the one of the most welcome way to establish the communication and after that once the communication is.

We want to maintain the quality of service over the network. So I found that some of the technical issues over that because when we I'm implementing the drone over that, as I said earlier, the energy is the major constraint. So under the heuristic algorithms, I'm not able to take much over space over the heuristic algorithm. So I can go for a meta heuristic algorithm and the metachemistic algorithm that I have taken is the Red Deer algorithm. Red Deer algorithm offers more. When compared to the heuristic algorithm, because each and every parameters when we talk about the quality of service, each and every term is dependent over the another term. So we need some maximum space in terms of exploration and exploitation so that I can search some more space. I can able to get much more space in the meta heuristic approach. So I have used the Reddit algorithm and once the communication is communication problems rectified by the. Heuristic algorithm. The second one I can take a metallic algorithm that is also OK for me. Next, I can go for enhancing this security. Security in this case I'm using the blockchain technology. Actually this blockchain technology is mainly implemented for financial transaction as we know already. But I'm using the blockchain technology over this network in order to enhance the security by simply implementing the neural network along with the Anopheles search algorith. And why? Because of the. So I spent the basic knowledge of blockchain technology. The third party is not able to enter without the without their permission of remaining existing blocks, which was handled by the members those who are in the already existing. For example, we have 3 number of uses. For example, I'm saying I'm controlling every, I'm controlling all the entire setup of a disaster network, but. Once unwanted people. Those who are intended to create a problem over that area again are not able to disturb those communication. What I have established over the disaster network to safeguard the people. So that is the main thing so far that purpose I have taken a consensus algorithm that is actually for proof of work and proof of state. So along with that I can add that up with some Anopheles search algorithm and to optimize the parameter and I'm getting. Maximum space over there and I can implement the quality of services. So these are all the three stages. The first owners are heuristic approach. That heuristic approach I have used the geography drone based communication. So in order to enhance the communication, second attack heuristic algorithm that is the red beer algorithm to improve the energy level once we are establishing our communication. And third one is the third one is for security For that I am combining the blockchain technology along with one more. Metabolistic algorithm that is nothing but search algorithm, that search algorithm can give a more workspace over that because each and every parameter in the quality of service is actually dependent on each and for example a packet delivery ratio or a throughput or energy or routing that is a major issue. So combining all those features. So we want to give us such a wonderful network for the disaster to manage the disaster and also we have to evacuate the peoples from those.

Area this is a major research contribution over there and in future I can establish the same research work that can be implemented with the artificial intelligence and also the the the recent trends like quantum computing. I can I can study over there to enhance my speed speed of transactions over the communications very important because that delay may affect May 1, we can be able to last one human being in the.

In order to enhance the communication or speed up the communication, I can. Now I'm studying over the quantum computing. That is the most interesting area for me. Two, we extend my research work from the base of my research professor.

Thank you, Professor. Your research contributions to disaster management and communication networks, particularly through the integration of drone-based systems, metaheuristic algorithms, and blockchain technology, are highly commendable. It's clear that your work addresses critical real-world problems. Finally, let’s discuss your industry engagement. Have you been involved in any consultancy or industry projects, and if so, could you share what you worked on and its impact?

Uh, actually, Sir, uh, I hope, uh, initiated now, but I'm not fully contributed over any industry consultation. And, uh, now I'm decided as of being a part of, uh, to help the farmer. Now I'm interested to do 1 research project over that. What's the project is, for example, a farmer is simply. Put all the seeds over there, then how the he's ready to that is the seed is ready to come out. Out of the soil and grown up and all over there. So due to the birds and the insights, the entire yield may get affected. So each and every time the farm is going there and making the birds to fly over or we want to prevent everything like that. So I can decide to analyze that what are all the factors that is purely affect the. Yields of a farmer and now I'm studying that bats is the major problem over there so. Thought affecting the bats, we want to create something some some some something to protect the yield or without affecting any living beings. That is the major objective of minds. Now I have started to do such a project to avoid the bats, which does not affect the yield of a farmer and I have visited more company but I'm not fully involved. So this is a true fact, but I have initiated the process to. Safeguard the farmers yields.

From the bats and other insects, Professor.

Understood, Professor. Thank you for your transparency and for sharing your aspirations to contribute to the agricultural sector with meaningful research. Your focus on creating solutions that balance both environmental and practical needs is admirable. With that, I have completed evaluating the key areas of your expertise. Thank you for your thoughtful inputs during this discussion. Is there anything else you would like to share or any questions you have for me?

Oh actually what are the expectations from me as being a professor to be get selected over?

So you're asking many questions to me. I I admire about that and if you want to play I can. Moreover I can add up some more few technical what I'm expecting over the face to face interview. There is any possibilities you can in what way I want to prepare.

More from this.

Thank you for your genuine interest in understanding the expectations and preparing further. As a professor, significant emphasis is placed on your ability to demonstrate research excellence, your proficiency in teaching complex material in simple and effective ways, and your contribution to both academic and societal advancements. Your ability to integrate cutting-edge technologies with real-world applications is certainly a strength. For future preparation: 1. Strengthen your narrative around how your teaching methods uniquely benefit students, focusing on how you inspire critical thinking and innovation. 2. Highlight specific, measurable impacts of your research contributions. For example, quantify how your disaster management solutions could improve response times or save resources. 3. If possible, engage with industry collaborations to demonstrate practical application and alignment of your expertise with current challenges. 4. Be ready to showcase how you can contribute to curriculum development, interdisciplinary research, and guiding students in both theory and application. Your enthusiasm and depth are evident. Keep building on these foundations, and you will grow into an even more impactful academic contributor. If there's nothing else, do let me know when you're ready to conclude this discussion.

Yes, Sir, I can follow. Thank you so much.

You're most welcome, Professor. Would you like me to conclude the interview at this point? Please confirm.

Our scheduled time for the interview is now up. Before we conclude, do you have any questions related to the role or the company?

No, Sir.