Hello. Yes, you can start.

Please start by introducing your professional journey in academia

OK, so. Very good morning, Good afternoon. I am Doctor Samiranjan Biswar. I have completed my B.Tech and M tech in electrical engineering from GI to Bhubaneswar and Ijit Sura.

Alright, so you’ve completed your B.Tech and M.Tech in electrical engineering from GI to Bhubaneswar and IJIT Sura. Go on.

And then after I have joined my. ESD in key DMT university. Also, I have completed my PhD in. Keep them to the university. My research area mainly focused on demand side management. Of a greenhouse using AI and IoT models. And also I have. Now I am working as an assistant professor in SIT Engineering College in the Department of Artificial Intelligence and Data Science. That's all about me.

Dr. Biswar, since you have experience teaching in the Department of Artificial Intelligence and Data Science, what is one foundational concept you most enjoy introducing to new students, and why does it resonate with you?

OK, so. I always try to. Introduce the. Practical oriented thing. Like how the artificial intelligence is related to our human brain.

So you like to draw parallels between AI and the human brain for your students. Can you give me a specific example of a classroom activity or analogy you've used to make that connection clear for beginners?

So it can the student can connect. Easily. I first actually make them understand that. Yes. OK, so just I teach them that. Their AI actually mainly 3 type of thing in like your. First one is your supervision. Supervise unsupervised and reinforcement learning. In there I actually make them understand that. How supervised learning is work like it? It is work like a teacher is teaching. Uh, all these, uh, to, to the, to the student. And like unsupervised learning. There is no teacher actually present. They have to cluster all these things according to the shape and size. And in case of reinforcement learning. Actually it's like that small. Heat. How he can actually know that this thing is danger after touching a, uh, hot thing? He feel that this this is actually harmful for me. So next time, he learned that from their mistake. And also. But if I will give some reward to that checked then.

Actually, he was motivated to do that type of thing, so I relate. With the reinforcement learning where I stays there. Yes.

So, you use the teacher-student and reward-punishment analogies to explain supervised and reinforcement learning — that’s a classic approach. Have you noticed any students struggling with these analogies, and if so,

So you tie reinforcement learning to motivation and real-life consequences for students. How do you adjust your teaching when a student just doesn’t connect with these analogies or seems confused?

OK so when actually student feel confused I first breakdown. My concept 2IN smaller part.

OK so when actually student feel confused I first breakdown. My concept 2IN smaller part. And take different different concept.

Dr. Biswal, I noticed your recent IEEE Journal paper, “Simplified Prediction-Based AI-IoT Model for Energy Management Scheme in Standalone PV Powered Greenhouse.” How has your research on AI-driven demand side management shaped the way you teach students about real-world applications of machine learning and IoT in the classroom?

Alright, so when students are confused, you break the concepts down into smaller parts and tackle each separately. Walk me through a recent class where you had to do this—what was the topic, and how did you adjust your explanation in real time?

And like I next I will go for some numerical. Like or code basic. OK, so when actually student got confused like in supervised learning. 1st I I actually gave example like teacher or student examples. If they still not understood that thing, I will go a little bit more example like. A student how give I will give a book. Through them. Then they have to read that. And after that, I will teach. I will ask that question from that book. And they have to answer from that book only. So in that case. If someone get failed.

Alright, so if a student still doesn't get it after your first example, you take it down another level—like giving them a book, having them study it, and then testing them only on that material. And if they fail, what do you do next to help them actually internalize the concept?

Then at that condition. The teacher have to retrain or reassessment all these things and. OK so I'm I am trying to actually get them. More number of example. Or individualized visualization tool. OK, so so that they can make more actually relatable thing to this concept. Hello, can you go for the next question?

Alright, let's switch gears a bit. Given your recent work in AI-driven demand-side management for greenhouses, where do you see the most promising, fundable research directions over the next three years? What kinds of grants or industry partnerships would you target to support that vision?

OK, so actually this AI thing as we know that this thing actually very rapidly evolving in every field.

Okay, so you see AI evolving in every field. For your own research area—AI and IoT for greenhouse energy management—what specific funding agencies or industry partners have you identified as realistic sources of grant support?

But I seen a that is very lag in case of in agriculture field. OK. So there are several project funding agencies there like I have recently applied in ICSS that actually funding from the agriculture and. Agricultural. Resource funding and DST is there one? One more is there? Anu Sundan is also there, so. A lot of agencies. Funding for this type of projects.

So, Dr. Biswal, let’s talk about your approach to teaching. Suppose you’re assigned a large undergraduate class—say, 200 students—and you’re not allowed to use slides or traditional lectures. How would you design a session to keep them actively engaged and really learning the material?

So, uh, for this type of uh. A large number of student. I will try to interactive class. And start with a real time real life problem. And ask the student to think and respond. Then I use. Well, the black blackboard. Or anything that, uh will present there. And I will make sure that small group discussion also or. Quick question, answer to student so that they can feel engaged there.

Right, so you’d kick off with a real-world problem, make it interactive, and use group discussion and quick Q and A to keep 200 students engaged—even without slides. How do you check if your approach is really working in that big a class, especially for the quieter students who might not speak up?

OK, so. First of all I will identify. Now which type of student actually not participating? And I will make sure that what is their real problem and I will discuss with them. Personally, 11 to 1. So that if they may feel lag behind about that concept. Then I will make sure that I will go little bit lower level. OK and also I will this I will give some extra tasks that simpler label. So that they can feel also engaged or get participate with their degree.

Alright, Dr. Biswal, let’s turn to something a bit different. 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 while I after joining that in. The department if I I have found that. There is some inconsistent. Then I will try to standardize the assessment method. And try to implement some new method that that have in implemented in big university like or in IIT or NIT. End I I'll try to. Follow a common uh format, uh for uh going forward.

You mentioned standardizing the assessment method and following formats used by IITs or NITs. But accreditation often requires faculty to collect evidence and demonstrate continuous improvement in learning outcomes. What specific day-to-day practices would you set up or monitor to make sure the department stays compliant with those requirements?

OK, so I would, uh. Makes uh, it's simple and routine like maintenance course file. Or with CNPO mapping Rubik's. And student work sample with each courses. Also, after each assessment, do a quick analysis. Note, uh, any. If if there any gap then I will note that. And record the all these small improvement. For the next cycle.

Okay, Dr. Biswal, let’s tackle a scenario around academic integrity. Imagine a colleague asks you to co-author a paper, but as you review their data, you notice some numbers just don’t add up. Meanwhile, there’s pressure to publish quickly for the department’s annual review. How do you handle that situation?

OK, so. As this thing is about research. I will try to. Actually, uh. Find out that uh. Uh, inconsistency. It if there is any inconsistency in data. And I will make them correct. Or else I don't agree to publish about that. And I will suggest rechecking or redoing this analysis.

OK, so. As this thing is about research. I will try to. Actually, uh. Find out that uh. Uh, inconsistency. It if there is any inconsistency in data. And I will make them correct. Or else I don't agree to publish about that. And I will suggest rechecking or redoing this analysis. Even if there is. Any traceries uh? On me.

Alright, Dr. Biswal, since VIT University puts a strong emphasis on industry collaboration, how would you bridge the gap between your academic research and real industrial applications in both your teaching and your research projects?

You made it clear you’d insist on rechecking the data, even under pressure—that’s good to hear. Now, thinking about VIT’s emphasis on industry collaboration, can you walk me through a time when you translated your research or academic work into something directly useful for an industry partner or company?

OK, so in this case I will align my research. With the real industry problem. Like I am doing now energy optimization or Iot-based monitoring and Ai-based monitoring system. And then collaborate with company for piloting project real data implementation. So that work moves beyond this only the simulator.

So for the next part, I'd like to focus on some of the core skills for this role. Let's start with image processing. Suppose you have a set of low-contrast images from a greenhouse monitoring system, and you need to automatically detect plant diseases—what preprocessing steps would you take before running a machine learning algorithm on those images?

OK, so I will first improve that image quality using contrast enhancement. Like, uh, histogram. Equalisation or a clay method? Then I will apply noise filter. End normalization. Followed by the sentiment to the isolate leaves before the feeding into the. My model.

Alright, so you’d use histogram equalization or CLAHE for contrast, then noise filtering, normalization, and segmentation before modeling—solid sequence. Now, thinking about embedded and communication systems, if your greenhouse sensor network starts dropping data packets during peak hours, how would you diagnose and fix that in the field?

OK, so in that case actually I will try to. Uh, first thing is that I will try to make like miss and uh. Network where the network can actually. Send multiple.

So you’d look at the network design first—maybe try a mesh or something similar so packets can take multiple paths. What about the communication protocol—would you consider changing that, or would you focus just on the physical network?

And in multiple parts. Uh, but the communication. OK, and I will also take the row. A signal strength in. Loss to deep, thick. Uh, I will switching the protocol like NQTT QoS or improving the uh network coverage. Hello. Can you go for the next? Answer question OK, so in in case of the better explanation in theory or OK, so in in case of the better explanation in theory or. Actually, I I will. Uh consented uh heavily on expanding the class part. The with the theory with an example.

And in multiple parts. Uh, but the communication. OK, and I will also take the row. A signal strength in. Loss to deep, thick. Uh, I will switching the protocol like NQTT QoS or improving the uh network coverage. Hello.

So you’d consider switching to something like MQTT with different Quality of Service levels, and also monitor signal strength and loss—got it. Let’s switch gears for a moment. When you’re teaching, how do you decide if a topic is better explained through theory in class, or through a practical lab session?

If it is applicable based on need or need hands on. Then I will. Get to the lab. Four student loan pay. Doing it.

You mentioned you focus on explaining in class with theory and examples, but sometimes practicals can drive the point home more clearly. What’s a topic from your teaching where you realized students needed the lab experience more than just lectures, and how did you adapt?

Okay, so you move to the lab if hands-on is needed. Can you give me a concrete example—a specific topic or experiment—where making that switch really helped students understand the concept?

OK, like one topic is IoT if I will teach IoT in only theoretical way. Then the student cannot understand actually. In a good way.

Right, just teaching IoT with lectures doesn't work well—students really need to handle devices and see the data flow in person. Switching gears, let’s talk about publishing. Can you walk me through the process you follow to select a journal for your research articles,

So that case I will do more more on handsome. OK, so in that case actually I will first. Map my journal with the my my article with the journal scope. And if there is several several similar papers are published then I will go for that paper also I then I'll check the indexing like Scopus or SCI. What is the impact factor? Review the time. Also follow the. Author Guidelines.

After that I will actually go for that. Of General.

Alright, so you check the journal’s scope,

Alright, so you map your article to the journal’s scope and make sure it fits before submitting. When you’re presenting complex research findings to a mixed audience, how do you structure your delivery to make sure both technical experts and non-specialists can follow your main points?

OK so first I will start with simple overview. So everyone get idea. Then gradually I get into the technical details. I also use example diagram and also adjust my explanation based on how they audience is responding.

Okay, so you start with a simple overview, then add technical detail, and adjust using diagrams or examples depending on the audience’s reaction. Now,

So that. Everyone can. In this same bus. OK, so to evaluate. The student actually what I have followed uh. In my class. Actually, uh. For some basic concept. I will. I will go with some basic and save some numerical or. Problem solving. And also I have gave some basic application based question so. After that I also align them with the course outcome. And use the clear marking scheme so that evaluator. Actually stay in.

Alright, you include basic concept questions, numericals, application-based ones, and align them with course outcomes using a clear marking scheme—that's a thoughtful process. Now, when guiding student projects or research, how do you strike a balance between giving direction and letting students take independent ownership of their work?

OK, so for the research purpose actually I give clear instruction, direction like. What is the problems to attend? What the scope and? I will go from the basic approach. After that I let them explore and make the decision. And I. Only when they are stuck. Or they need some help? At that condition I give some suggestion to them.

Alright, so you start with clear direction on the problem and scope, then give students freedom to make decisions, stepping in only if they get stuck. For this role, having a PhD in a relevant specialization is critical—can you briefly describe your research focus during your doctoral work and how it connects to the areas you'd be teaching or guiding here?

OK so my research mainly area is AI and IoT implementation.

Makes sense—staying updated is so important, especially with how fast AI and IoT are moving. Alright, we're wrapping up. Is there anything you'd like to ask about the role or the team?

In the in a smart. Where I am trying to solve the energy management. And I am trying to. OK, so I want to know that what type of role I should get there?