Good day, Dr. Debasis Acharya. 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?

Please start by introducing your professional journey in academia.

Hi, I am Doctor Devasis Acharya. I have completed my PhD in control systems in 2024. From then uh yeah I am involved in teaching artificial intelligence, machine learning this subjects in exc department. I have compared two years experience in computer science, artificial data science and cybersecurity department. Last two years I have taught students.

Artificial intelligence, machine learning and data analytics, robotics. This subjects generally I teach with simple simple example of real world and then start the theory and mathematical concepts requiring machine learning, artificial intelligence etcetera.

I have also here in all many students for research work work where the studies students are involved for data driving control design for different fields like artificial ventilator, knee systems, knee systems and one of my students get best IEEE conference paper except paper accepted.

Also as few students have also submitted articles in machine learning related subjects. My two my research. My academic background is in electrical engineering but now I have integrated modern 21st century based technology like artificial intelligence, data science with my knowledge and solving real time problems. Even as I like to teach with.

With real examples, so I invoke.

Students for project Beslani that means like in present time I am teaching at IoT. In this case I also involve students with a very first day with simple. I do this example and I try to gradually make interest between students and then I introduce about the related theories and for that it clears their knowledge also they feel happy to. Work with me.

I also have few M tech students in during my PhD in NI National Institute of Technology relevant. I also have guided on behalf of my supervisor. I guided few scholars and tools have also published research articles in my academic journey. I have presently more than 18 sci channels. I have I triple E transactions and.

Location in nature scientific report by fuzzy logic during the PhD is different from the from the traditional method where I have used optimization algorithms to tune their settings like say rules etcetera. So I feel very happy when I see my academic journey because now I am able to apply. Modern technology based software. Computing based or physiologic based concept in my research work where I integrate my electrical engineering knowledge with the modern technology based knowledge that is the artificial intelligence, machine learning and make different research papers. Presently I have mainly focused on product based learning. I try to. To integrate students with product this work whether it may be a simple product.

But the product that will keep them the real experience and like one of my students have recently paid a smart mask using ordinary Ordino kits and available fast and it is Kate special prize in a national level science exhibition.

Thank you.

Thank you, Professor. Let us now focus on structured evaluation. I’ll begin by verifying specific skills relevant to this role. Can you elaborate on your expertise in Artificial Intelligence, Machine Learning, and Data Science? Specifically, how have you applied these areas in your research or teaching practices?

Presently those by past uh research work was based on traditional control logic, but after a uh, after letters, I use soft computing bits technique in my research work. Like in my career, I have more than many people, more than 10 or 12 papers, not only 10 more many papers where I have used meta heuristic, this algorithm for tuning fuzzy.

Settings like rules, save, etcetera for and I have applied it for artificial. For artificial ventilator system I have mostly focused on on different type of development of metaheuristic algorithms. Presently I am using machine learning techniques like Q learning or reinforcement learning which is a environment feedback. Model where? Where I have which I have used for artificial ventilator systems and even me exoskeletal systems for design is designing a controller. But I believe that artificial machine learning technique which is a data-driven technique can perform better for control design if we integrate these techniques with traditional method I have. I have not used directly AQ learning method or any. The field learning method directly for designing the controller but I have used them to tune traditional controllers parameter like PID info, PID controllers and designed many control systems for ventilator systems. Even these things I introduced to students such that they can understand how they can use this concept in making any.

That's what. So data-driven control system is very important which is a which is where that the control logic will can take the decision by itself. But here we are not directly controlling the any system like artificial ventilator or knee system. I use the machine learning techniques in a different way to tune the controller parameters.

Which keeps the very good response and that I have published in a channel and I I have more than few words which I already have communicated in different group channels for publication in artificial intelligence. Which I feel is very important and very interesting field for data-driven control design. Because control design is itself a complex because most of system in the. What are very nonlinear, where traditional methods fails to uh control the system because it needs mathematical model of the system, which is very difficult to get for a for for a system like biological system or biomedical devices or a complex system. So we cannot use traditional controller or we feel very difficult to design a controller accurate controller.

Using such traditional method but here if we can use the artificial intelligence or data-driven concept, machine learning concept which is a part of AI.

For designing which will not involve the mathematical which is not required the mathematical model of the complex system because it can can communicate with with the environment of the complex system and can get decision for tuning a parameter of a traditional controller like PIT or IFOPID if in many nonlinear control design like sliding mode.

Control etc. We can use the machine learning model machine learning model to tune them. Even I recently I have worked one neural controller where I have used neural network concept. Simple concept of taking input with weighted sum of the output we take activation function. This concept I use and I make it like a pit controller where I try to.

The negative side of the π traditional proportional derivative, integral controller that is the pit controller like derivative and integral action in the PID controller. Instead of this action I have used the neural networks such that it can behave like a PID. But the problem is that when I change the system settings then I am not getting the correct answer. That means I feel that machine learning.

Can be used or not can be used for such a such complex system regulators. But if I not use directly instead of that I can use it for getting the optimal setting of the PID controller or any nonlinear controller. Then it gives the correct answer. Thank you.

Thank you, Professor. You've demonstrated a clear understanding of incorporating Artificial Intelligence and Machine Learning into control systems and their applications in both research and teaching. Moving to the next skill verification: Can you explain your approach towards teaching theory and laboratory courses? Specifically, how do you ensure students grasp both the theoretical and practical aspects effectively?

What I feel during my teaching students need a curiosity. If I can generate curiosity between among the students, that definitely will learn. And I try to make that curiosity with some simple example first. Like in data analytics or data analysis subjects, Data science subject, I first explain different real life examples.

Then gradually introduce the concept and make them curious. They ask me what is the next what is that? Then I explain the things in case of laboratory or lab also like machine learning lab or artificial intelligence lab first the real first with the real example. When I explain the concept then and give them simple way to solve the problem using the Python. Libraries, but not exactly the Python libraries the first into. I first introduced them to do the work by the simple mathematical equations or simple things like. Suppose for a data science subject they need to find a main, median, etcetera. They will not accept for very first time. They will not use the any package. Instead of that they will use the general mathematical. Rules or formulas to find that?

After till gradually I introduced him the libraries. I believe that if they solve the problem by their very simple knowledge then they can easily understand it. After that if I introduce them the libraries required for solving different problem they can easily do it like machine learning linear regression. In linear regression different packages are available but first I introduce them the linear.

And I tell them to introduce to use the equation for making your model. After that they use so. So for teaching the the subjects I generally use very real example with simple explanation after thing introducing a example with very simple example with a Python or required tools and then gradually go in depth. And in present time I feel very.

Interesting in ChatGPT or artificial intelligence based applications because they can easily create very simple example which is understandable by the students. I sometimes use that, use that and as in my research also I use sometimes to get some concept or explanation which later I read again and again and with my experience I try to. Try to get the exact result.

So whatever I'm getting from that strategy or DI tools, then I introduce it in my research for teaching. Also, sometimes I use AI tools, ChatGPT, etcetera. For formula, for structure by problems, whatever I want to explain, I first structure them and then explain it. Thank you.

Thank you, Professor. Your approach to blending foundational knowledge with practical applications, complemented by AI tools, is thoughtful and demonstrates a structured teaching strategy. Next, regarding your experience in student evaluation and exam duties, could you describe how you assess students' understanding effectively while maintaining fairness and academic standards?

Whenever I evaluate students marks generally like there are two parts, 1 is internal, another is suppose external, internal means internal marks that is that added from their assignment, their presentation, presentations for a subject, topics and they are internal assessment etcetera and external that is below that is for the intern examination.

So during the internal assessment, I see they are writing their knowledge. I don't want to see, I want to find the many sentences they will write but I want to I check whether their concept is clear or not. Even during the Viber or or interaction with the students, sometimes they ask questions but that is a simple I mean my aim is to judge whether their concept is clear or not.

I don't, uh, feel that, uh, if they can remember or memorize the fact and explain me like a book, uh, etcetera. Uh, yeah, it is the fruitful. I feel the fruitful thing is their concept. So I check it in their writing also. And if I get satisfied, I keep them good marks. And because marks is also important for them and due for the external examination, the evolution is like that. What the content?

What is the what they are writing is very important for assessment for getting a proper assessment and during as I am working in a university so I can prepare the questions for their final exam. During the questions I also focus these three things. Basically one is conceptual questions where I judge I want to check their concept is clear or not. Second is.

Heart heart question that needs their not their knowledge or experience in the subject such that it can they can explain it during the writing these things and another thing is that little bit of example type or such that I want to judge that is thick and you rather they can use some mathematical concept or not some exam real example or not etcetera. By this I.

I interviewed.

Thank you.

Thank you, Professor. Your emphasis on conceptual clarity, alongside a structured evaluation approach combining assignments, internal assessments, and well-thought-out exam questions, clearly aligns with academic standards. Shifting focus, can you elaborate on your ability to guide student projects and research? How do you help them define research problems and support their progress throughout the project?

This is the main things what I always do during my teaching. Generally from my very first day after few talking with the students, I generally try to motivate them such that they can do something. In reality they can involve in research work or project based because few students are there those are interested in writing some things or research oriented works but few students are there also. Do those are very happy to do.

Yes.

Do something we we hardware or software, anything. So I first try to motivate them for both cases such that they come to me and involve themselves for a research work or a project test work. And I give such students also. So generally when any students come to come to me for a research work or a project based work, I first give them the opportunity to find their interest.

What in which field they feel interest or what they are expecting to do in real world? Sorry, real world.

If any students cannot give a clearview and he or she needs help from me, they needs my suggestion, then I give some many two to three or few topics or interested areas where which I can help them and they can get interest in my student. In my case from the beginning I have different type of students, few students those are getting. Their interested works and.

From me and to some some few students, those are coming to me with their interested topics. I help them to build a problem formulation, how I guide them, how they will move on and and make it. If it is a research work, then first I try to make them about the concept, then how the to write the paper and gradually submission etcetera if it is a product.

Gradually one by one we build the product. Like I have helped one students for making eco sutra type was type apps which is a for for waste management. Another is for making a ordinal based smart mask.

Another is called like this, I, I help them such type of product based, uh, product or project based work and help they guide few students for, uh, for research type of words. And in research, mostly I try to, if the students is getting interest for AI, machine learning and deteriorate control-related work, then I, I help them in that way if they have some different. Knowledge or different concept which is actually not beyond of my knowledge.

Then I try to learn it and help them and I have done one work also like one students D-Day keep 2 currency cryptographic work using meta heuristic where where I helped him to guide for the meta heuristic and such that he can use it for cryptographic technique actually which is not in my research work but I adapt myself to learn it and then I teach him or guide him for the research work.

Thank you.

Thank you, Professor. Your dedication to fostering both individual and collaborative learning, along with your adaptability in exploring unfamiliar research areas to guide students effectively, is commendable. Let us now assess your communication and structured teaching approach. How do you ensure clarity and engagement in your lectures, and what methods do you employ to make complex topics accessible to a diverse classroom audience?

Like if I start from the complex topics like many students have a concept that machine learning means a mathematical things, mathematical foundation, they should know different maths. And it is true that many students generally afraid of maths. So I never start with the maths, I always start with a simple real example.

Concept based and when it is required to introduce the maths.

Then I introduce it gradually like linear, linear, linear equation or etcetera. I never start with the mathematics first. I go in the simple examples then. So for engaging students, the first thing is that we need to make them curious. I never believe that from the very first day or beginning we should teach complex theory or the many concepts.

I always believed to start with motivate them with real example with their little bits of their career guidance. Why they will read it, What is the requirement of studying artificial intelligence, machine learning, sort their subjects like electrical engineering etcetera. They know their parents or their familiar ones have told them to come here or they have guided from the others and join the course maybe.

But I I feel that we need to make them first curious. That curiosity definitely will drive them to teach Yes, every day. It is not possible that they will engage effectively. Maybe their brain will get in different direction because of their age. But with the simple example, with the little bit of hardware or little bit of a practical example, I can I always try to increase them.

In during my class.

Thank you.

Thank you, Professor. Your approach of starting with relatable examples and gradually introducing mathematical foundations showcases an excellent balance between accessibility and rigor, ensuring inclusivity for diverse learner profiles. Next, could you discuss any significant industry projects or consultancy experiences you’ve had? How have they influenced your academic journey, particularly in linking practical applications to theoretical teaching?

As I have experienced in high golden directly industries and in my research areas, I use that experience. Experience is is the time completed or I can continue.

So.

Thank you. So as I have more than four years experience in industries and I know in theoretical concepts or whatever we are studying during our graduation is little bit of difference from the practical during my study or my engineering. I also feel something different when I join in industry. So as I have that experience, I try to whenever I give any example like.

Recently I'm still teaching data science students about machine learning, artificial intelligence and robotics. Also I first I sometimes use the industrial things industrial needs like suppose in case of machine learning, many people say about the Flipkart Amazon example. I also give them example like here like temperature monitoring of. Heat of the steel plant or in a in a pharmacy, suppose a medicine is. In they are doing research in medicine they can't they monitor different different.

Information regarding a chemical for testing information, how they use this, details for making a model like this. So I generally use industrial, industrial knowledge or experience with that is a real example with the students to make them understand of different things.

Thank you.

Thank you, Professor. Let’s now discuss your research contributions. Can you summarize the impact and significance of your research publications in reputed journals, particularly how they align with the advancements in Artificial Intelligence, Machine Learning, or related fields?

So my PhD work basically was on meta heuristic algorithm based control design for artificial ventilators where I designed many meta meta heuristic algorithms. Basically I have used class tupper optimization algorithm developed by my supervisor and I modified it for making the control design tuning the PID controller for ventilator. I also have developed one algorithm that.

Is the Human Consistent Based optimizer which is published in Nature Scientific Report report which I use for tuning the PID controller. If you see my research work you will see I have many publication for fuzzy logic control but it is not basically the I have not used the available concept. I have developed a method. Which by which I have tuned the setting that is the roles save for the fuzzy logy and I.

Use it for tuning the PID controller to make it like a online based controller where the fuzzy logic is a knowledge base system where I have tuned to optimize their settings. And this optimized fuzzy system is used for tuning the PID controller whenever any changes in the ventilators like patient change or lung system change etcetera or any noise it can automatically get.

Ideologically keep the PID parameters for the ventilator system. So if you see my if you, if I if you see my publications, you will see I have used the important technology that is the soft completing best technology with the traditional method like a PID controller for the research works normally. Presently I have.

18 HCI channels I have publications in IEEE Transaction One, IEEE Transaction.

To journal in scientific report nature many else fears channels preacher nature channels publications I have I have more than 16 I triple E conference. 18 CI channels, 7 Scopus publications and all are basically Q1 or Q2, not even Q3 or et cetera. I have Q1Q2 publications, the highest 8.6 impact factor. My present age index is.

11/18 is 13 and uh, present uh, citation is 378 like this. And I my my other is my research area is the first area also because my I have worked in artificial ventilator system for my PhD. But with my interest I have publications on power system also where I have used such type of uh.

Based in better way, better heuristic algorithms or soft computing techniques for relay coordination problems. Even I have a publication in nuclear model. It is true that these things are not machine learning. Meta heuristic is not machine learning, Fuzzy logic is not machine learning. But this is 1 approach of a learning. Where subcommittee logic is a soft computing techniques knowledge based system. This is not like directly not a data pattern recognize like a machine learning tool. Where machine learning techniques generally read titles and find the pattern of the data then get the decision or do something better. In fuzzy logic it is not like that, but it is a soft computing technique which is not.

Time