Thank you.

Please start by introducing your professional journey in academia.

Good evening and thank you for this opportunity. My name is Rajagopal Pesapati and I have completed my PhD from National Institute of Technology, Jamshedpur.

My doctoral studies focus on power system related.

Problems with solution with artificial intelligence where I integrated machine learning techniques and data-driven techniques to address complex engineering problems.

My current research work includes roles of.

Artificial intelligence, machine learning, and deep learning approaches to solve.

Complex electrical engineering problems. Recently I have applied deep learning technique.

Namely gated recurrent units to forecast and classify the.

I'll market clearing price in mainland Spain and Australian electricity markets.

Focus completely on data-driven techniques which are quite easy to implement for the power system related problems.

My current research work.

Also previously I have 12 research journal papers in reputed journals and.

Which shows my ability to work hard and.

High quality, constant. In terms of teaching, I have hands on experience as a course instructor to the core electrical engineering subjects like control systems and power systems.

Also.

I have submitted it.

Few projects based on.

Data-driven techniques to the research promotion scheme supported by the Act.

Very soon I'm going to attend the phase two evaluation for the RPS submitted project supported by the AACTE.

This I can say that.

And eligible and.

Working hard for the.

I think artificial intelligence and machine learning professor at VIT University. Thank you.

Thank you, Professor Rajagopal, for sharing your academic journey and research expertise with us.

Dr. Rajagopal Peesapati, I have reviewed your credentials and the job requirements. Let's proceed with the skills verification related to the Professor role specializing in Artificial Intelligence and Machine Learning. I will ask specific questions to assess your expertise, teaching abilities, and other key aspects outlined in the position. We'll begin by validating your expertise in Artificial Intelligence, Machine Learning, and Data Science. Could you explain how you’ve applied deep learning techniques in your research projects, specifically in relation to the state of charge (SoC) estimation for electric vehicles?

Yes, thank you. Uh, I have here.

Huge data.

Which is having the which is also called as the historical data of a battery of several electrical vehicles like it may be a car, it may be a 2 Wheeler or it may be a public transport like buses or some.

Which are used in the which are used as a goods carrier. So all these data has been.

Other electric, electrical, commercial vehicles.

Consolidated in one form under different circumstances. Consolidated into one form and separated as a training and testing data.

So the training data was given to the.

For example, if you assume the deep learning technique as a gated recurrent unit.

That will be applied. The training data will be given to the gated recurrent units.

Actually before giving to the greater recurrent units the data has been pre processed or it can be said as normalized so that the data.

The evaluation will be easy and it will be very.

Uh, perfectly tuned with the gate recurrent units during the period of training.

Other the testing data will be utilized to test the.

Already tuned Gru parameters.

Already tuned Gru to uh to test uh. The testing data will be given to the already tuned Gru to obtain the outputs and these outputs have been compared with the actual outputs so that the errors will be found out. Their errors like percentage mean absolute percentage error.

And mean absolute error, so these can be utilized to.

Root mean square error.

Find the.

Appropriateness of the proposed approach.

So here the state of charge is an important thing in electrical vehicles which can be used to determine how much percentage is still remaining or how much percentage has been consumed. So the correct estimation of the state of charge will be useful for a particular user like either it may be the owner of a 2 Wheeler, owner of a four Wheeler or.

Are a commercial electrical vehicle based commercial vehicle. So using gated recurrent units are a deep learning approach in order to obtain. If you obtain this correct prediction, it will lead to the.

Perfection in the OR the long using long lasting use of the Surface battery.

In this way, we apply the deep learning approach to estimate the state of charge of a electrical vehicle.

Thank you.

Got it. Based on your explanation, it’s evident that you've applied deep learning techniques like gated recurrent units effectively in your research. Let’s move to the next competency. How do you approach teaching both theoretical concepts and laboratory sessions for students in emerging areas like Artificial Intelligence and Machine Learning? Can you provide an example of a methodology you use to ensure students gain both conceptual understanding and practical skills?

As for sure, I will let you know how to incorporate the laboratory sections or I can give a reliable example so that the students can realize how this machine learning and artificial intelligence can be utilized in the laboratories in general.

The regional load dispatch centers, what we call as in a deregulated market, it is a called, it is called an independent system operator. These are utilized to.

Obtain the market clearing price a day ahead before.

24 hours a day ahead before the uh exact day to predict. So here the electricity market clearing price is obtained from the bidding strategies of both generation companies and the distribution companies, which is a direct relation in the real world problem of the electrical engineering.

So if suppose if this.

Any company who is participating in this pool based electricity markets, if they came to know this is the electricity market clearing price for the next hour or the day ahead. So they if they bid at that with that price then they come under or their distribution comes under the this pool based markets. Otherwise they may not be selected.

So if they have not selected, they cost of delivering will be very, very high. So the maybe the energy consumed or the price at which the consumers receive will be very, very high. So for a company either it may be the distribution company or a generation company. Those who are participating in this bit must forecast this electricity market clearing price at a.

A day ahead before at a very high accurate.

Algorithms. So in this place of algorithms, this machine learning and artificial intelligence algorithms will play an very, very important role. Recently, I have experimented the same thing in forecasting this market clearing prices or what we can say electricity prices for two different markets, namely mainland Spain and.

Mainline Spain and Australian electricity markets. Simply I have collected so much of previous data and I have divided into two parts like training and testing and I have implemented the deep learning approach to obtain the exact market clearing price. I have received a very good result and the results have been submitted to I triple transactions on industrial informatics which for which I have received the major.

And I am yet to submit the revisions very soon.

So in this way in the laboratory sessions I will give these type of examples.

To come to know how to how this machine learning and artificial intelligence techniques will receive an intelligent approach or a basic benefiting approaches in laboratory sessions for electrical engineering-related problems. Thank you.

Thank you for the detailed explanation. You've demonstrated a practical and research-backed methodology to bridge theoretical concepts with real-world, hands-on applications effectively in laboratory sessions. That's a critical skill for teaching in this domain. Next, let's focus on guiding student projects and research. Could you describe your approach to mentoring students, particularly in helping them define feasible research scopes, select methodologies, and prepare for publication?

OK, sure, I will guide the students in.

From mathematical based approaches, so I would like to assist the students.

Like first of all, when anyone need to search or do a research good research work, they need to go with standards of the research work. So I will assist them in reaching a good standard journals so that if the standards are good, if their foundation is good, the output will be automatically good. So. I assist the students in in reaching their foundation.

In reaching certain goals.

Bear in a good.

Good conceptual papers so that.

The research will be very, very good. The research outputs will be very, very good also. I let them work. The innovations are currently in the progress.

Or in in the field of market in regards to the artificial intelligence and machine learning so that every student.

We'll try to apply the artificial intelligence and machine learning problems, where the feasibility is there to the electrical engineering problem, students electrical engineering problems also.

I have hands-on expertise on coding this machine learning and artificial intelligence algorithms so I can help them in.

So in this way I will give a step by step procedures to find or to get a good outputs from the.

If there is any problem with their coding related skills.

Students by guiding them at every stages at every stage. Thank you.

Understood. Your response highlights a structured and supportive approach to mentoring, focusing on strong foundational research, current innovations, and practical guidance in areas like coding. This seems to align well with guiding students effectively in both projects and research. Next, let's explore your ability to evaluate students and manage exam-related responsibilities. Can you discuss your experience or methods for assessing student performance, ensuring fairness and consistency in their evaluations?

Yes, of course. I will.

I will guide the students during the my teaching classes only. I will guide the students to follow certain ethics like.

Solving a problem solving a problem on their own even though the problem is has been written or has been solved on the board so that the student will improve their or announce their self learning mechanism which will turn them to be an ethical person and not to deviate from this ethics. So the same pattern if the students who were trained in this manner may not give any.

Misconduct during the exams at the At the same time, I will put my 100% efforts in teaching as well as the in the evaluation of the scripts like during the evaluation, I will make a.

Key structure of the question paper or key answers, answers answers script.

So according to this I will evaluate each and every.

Every script of the student so that none of the student will be highly biased or lowly biased. Each students will get equal prioritize based upon their written skills. So I hope I I believe this type of structure during the teaching and during the evaluation which make the student bright and and ethical person.

Thank you.

Thank you, Professor. Your emphasis on instilling ethics, encouraging self-learning, and ensuring a detailed and unbiased evaluation process reflects a thoughtful approach to student assessments. Moving on, how do you ensure clear communication and structured delivery in your teaching to engage students effectively, especially when dealing with complex topics in Artificial Intelligence and Machine Learning?

So every artificial intelligence technique or a machine learning model or a deep learning model have their own basic their own concept. So whenever a student.

Yes.

A student should learn if If a student would like to learn, they must show first what is the interest, how they have to we have to grasp. I grasp their interest. So for that, first of all, I will give my 100% effort in learning several algorithms and several applications which are really which are very.

Is it very real in the present fields? So whenever I give this type of examples where this artificial intelligence and machine learning techniques have been utilized?

Automatically the student pays attention towards the core. How to know how? What are these? What are, how to know what is the artificial intelligence? What is the machine learning? What is the deep learning?

This is my step one Step 2 I prepare many seminars, PPTS and I will use so many number of tools so that.

Conventional method of teaching using blackboard or chalk and chalk and talk will be.

I may not using the chalk and talk, I may not provide a complex diagram perfectly on the board, but utilization of this seminar and PPT I can provide a complex diagram and explain the students up to a.

Replaced with some PPTS so that the student will not get bored at that same time he will learn. I can give.

Up to so much of core level. So maybe these two will provide a student to grasp these two will provide will grab the student detection towards in teaching and teaching methodologies. Thank you.

Thank you, Professor. Your focus on incorporating real-world examples, modern teaching tools, and dynamic delivery methods like seminars and presentations shows a strong commitment to engaging students and simplifying complex topics effectively. Next, let us touch on your experience with publishing academic research. Could you detail your approach to getting your research published in reputed journals, particularly your strategies for choosing suitable journals and preparing manuscripts?

Yes.

Actually I have.

I hope.

Very sound experience in publishing research papers in highly reputed journals like I Triple E.

And in that also I have top rated journals like. When it comes to the choosing a journal it is very much important to which concept to which concept I am choosing this journals like.

And elsewhere.

I need to 1st confine to a research work Oregon. I need to find a research gap.

Based upon their research gap, the RA I will.

Done some research work and obtained some results.

And these have been abstracted in the paper and research paper. And this abstracted one I will find a suitable journal based upon this abstract. Many journals you utilize as a journal Finder that is suitable to your research work. But most of the journals does not meet these requirements. So I prepare a plan like I select few journals.

In a priority wise so that once if I submit a paper to the journal.

In a.

It must not returned before the review.

So this is my obviously this is my the strategic plan. It is not a strategic plan, but.

Came I if I need to submit a for a journal, I will search that topic which is related to mine and I start preparing the research work. So in this way I communicate to the journals and most of the journals accept my research work as they are directly.

Are directly related to.

This facts will be showing my interest to the reputed journal reputed journals. Thank you.

My research concepts answer this.

Yes.

Please go ahead, Professor.

Yes.

Actually I am working presently on. I have my research lab comprises of OP means real time simulator and I have a power quality analyzer also.

So the nearby research institutes approach me for getting the real time results in using for the MATLAB Simulink files using the Opel RT.

So also some of the industries which are based upon this power converters.

To the nearby institutions. To my nearby my institution.

It'll approaches me to do some real time simulations so that hardware will be designed based upon these ones. So.

One I have popularity real time simulator as well as power quality analyzer will be utilized by recent uh nearby distribution companies to check their power quality ratings as well as during the energy audits. The power quality analyzer is highly utilized for nearby industries like scalp industry.

So I am dealing with presently these two equipment which are under the consultancy projects at my institution.

Are wood industry and some battery management system industries.

Thank you.

Thank you, Professor. Your involvement with industry-focused projects, including real-time simulations and power quality analyses, highlights your practical expertise and significant contributions to applied research and consultancy. Based on the key responsibilities and qualifications for this role, we’ve gone through several crucial areas. Do you feel there’s anything specific about your background, expertise, or approach that we haven’t covered but you’d like to emphasize?

Yes, I would like to say that I am very hardworking person and I commit to the ethical discipline and I'm very punctual in nature.

So this is my behavior behavior wise and during my post graduation.

At NIT Jamshedpur National Institute of Technology, Jamshedpur, I have received a silver medal for showing the academic excellence in the year 2011.

So I can say that.

I will. I am a hard working person with a.

Strong working skills at Vellore Institute of Technology.

Hello.

Thank you.

Yes, I would like to proceed.

You're welcome, Professor. That concludes the skills verification process. If you’re ready, we can move to the closing part of the interview. Would you like to proceed?

Professor, to begin, could you share which particular course or subject in artificial intelligence and machine learning you find especially fulfilling to teach, and why?

I would like.

Dr. Rajagopal Peesapati, I see your research portfolio includes work on critical topics like state-of-charge prediction in electric vehicles using Bi-LSTM and optimal charging coordination in electric vehicle systems. Let’s start with this: you worked on "State-of-Charge (SoC) prediction using Bi-LSTM." Can you explain how this specific research contributes to both the field of machine learning and its practical applications, particularly in electric vehicles?

Yes. So the buy LSTM bidirectional long short term memory network, this is a deep learning network.

Which has many number of gates and many number of reset signals also.

Excuse me?

I have collected the data regarding the state of present state of church.

Excuse.

Certainly, Professor. Please take your time to clarify or elaborate.

Yes, thank you.

I have. I have continued. I have.

I see—feel free to gather your thoughts. Could you continue explaining how you used Bi-LSTM and the data you collected for state-of-charge prediction?

Yes, I have collected the state of charge.

For the present time and temperature settings of the battery as well as the output voltage of the battery as the key parameters to find the next level of the state of charge by using by LSTM.

So this data has been collected for near about 10,000 vehicles which are having two Wheelers.

4 Wheelers and some commercial vehicles, so this data is directly available in.

User interfaces and Internet browsers and so many number of.

Websites. So this data has been collected and divided into two parts as Part 1 is a testing training and another another part is training sorry testing. So when this bias team has been utilized it is.

The by LSTM code has been deployed in or developed in Python programming Python language and I have used skillet learning program modules to obtain this LSTM modules. So once I have used this LSTM modules I have given the input from an input file and this input has been preprocessed. Or normalized between.

10 and 1.

To learn the, to learn and to to train the by LSTM at a easy at a high accuracy. And finally I have tested with the same data with a reduced dimension to obtain the exact.

Actual output values. So finally I compared the obtained result with the true values to obtain the different errors. Thank you.

Thank you, Professor. You've outlined the technical approach quite well. Now, as a follow-up, could you explain how you would convey this complex concept—Bi-LSTM for state-of-charge prediction—to undergraduate students in a classroom setting? How would you simplify it for their level of understanding?

Yes, it is. It is a little bit calm, uh difficult, but I will try my level best.

Whenever I give this before going for the LSTM, I profoundly teach the students what is the need of.

Forecasting the state of charge in the next time interval or the future coming interval.

So if I able to.

Uh, and reach the students, uh, perfectly. Then it is very easy to, uh, explain the by LSTM, how it will evaluate it easily. Actually, before going for the by LSTM, I will give a best example of neural networks, which is a very, very simple neural networks, artificial neural networks, which is a very easy topics in understanding at the undergraduate level. So when.

If I able to discuss the same thing.