Good day, Elakkiya Prakasam. 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.

OK, uh, hi, I'm Melaka Dr. Ielaka Prakasham. Uh, I'm a high performance computing professional, uh, with over 10 years of experience in, uh, academia, research and as well as in industry. Currently I'm working in, uh, giant pine solutions, uh, which is associated with the Turing, uh, where, uh, I'm working in. Validating the LLM generated QR code and how to improvise the performance of the.

Uh, code in the Gen. AI project, which is where my client is NVIDIA, uh, and also I'm working in, uh, Amazon's agentic AI, uh, workflows. And previously in my PhD, I worked on, uh, cache efficient parallel bloom filter architecture, uh, which is, uh, in using open MP in multi core architecture. And during my academic journey in vit, I worked as an assistant professor where I.

Handled number of laps including digital design and Python And also I assisted students in mini projects and as well as the projects.

Also.

Umm and uh, I'm very much interested in teaching and mentoring students and as well as I want to pursue my research in VIT along with the institutional goals.

This is my, uh, short feed, uh, short intro about myself.

Yes, yes, sure.

Considering your experience in guiding student projects and teaching courses, how do you ensure that students of varying abilities and learning paces benefit maximally from your courses?

Actually, uh, my students, uh, uh, will be, uh, first, the will be teaching students with a strong fundamentals. So students will be varying their learning. Learning paces will be, uh, different considering with the students. So by uh, with the help of assignments, we can, uh, we can evaluate students performance where, whether they are lagging or whether they are proficient.

With the help of business assignments and as well as the projects we can evaluate the students performance and in the labs with the experiments how they are doing. With the careful monitoring of the students during the experiments, we can understand their learning and their absorb how they understand the concepts clearly.

Yes.

How do you adapt your teaching methods when students struggle to grasp core concepts, despite these evaluations?

So some students obviously in a group of 40, group of 40 or group of in a class, some students will be lagging to understand certain concepts. So we need to give some more importance to the students in understanding in while explaining a concept to a class. We can't we can't come to a low level and explain but for some low performance or some students struggle to understand the concepts we should be.

Giving a real world example by with the help of real world examples and we should give some extra care to the students. In understanding the concepts through some real-world examples or with the different.

Relate. By relating the theoretical concepts with the practical or hands-on assignments or examples, we can make them understand more clearly the concepts.

Yes.

Can you elaborate on your experience with guiding student projects and research, particularly any innovative or impactful outcomes you've achieved?

In my handheld lab with AI and as well as in digital logic design, even I handle the labs with digital logic design. Even the some students have struggled in understanding how we can map the logic with the use of K maps itself. So we personally.

Actually.

Engage the students to elaborate the. Logic. Boolean logic. How it maps to the?

Hardware and how we can, umm, uh, uh, realize the, uh, Boolean logic with the hardware through the simulations, how it is happening. We, we need to give extra care to the students and there we could see, uh, some improvements in understanding with the, uh, experiment and theoretical with the hardware level.

Focusing on hands-on understanding of hardware and simulations is an excellent approach. Let's discuss your research experience. Could you explain one of your key research contributions and its potential impact?

Yes, actually in, uh, during my research, I focused on my high performance computing. So I designed a cache efficient data structure that is a bloom filter data structure. So first way I analyzed the, uh, bloom filter architecture there I could see them some uh, misses in the locality of cache locality. So what I did is I improvised the cache locality of reference by assigning. By accessing the memory location based on the cache line size.

Whereby I improvise the cache efficient. I could achieve the cache efficient bloom filter and also I parallelize the insertion on the querying of data in the bloom filter. That is a data structure where for that I used a large DNA data in the application of bioinformatics application which is. Chema sequencing Chema search in the DNA came came a search in the DNA sequencing applications.

There I could achieve a high performance as well as in the parallel, uh, through parallel and as well as the cache efficient uh, approaches, I could achieve more, uh, performance improvement in this bioinformatics application. This is my, uh, broad research area.

Designing a cache-efficient Bloom filter and optimizing it for bioinformatics applications is indeed a significant contribution. How do you envision integrating such advanced research concepts into your teaching methodology for graduate or undergraduate students?

Yes, actually during my research I found a very interesting actually that time only I I got to know that in the CPU itself we can achieve that much parallelism, massive parallelism. So it's not a big deal to nowadays under teaching the undergraduate students how to exploit the multi course available in the CPU with the help of the languages such as open MP, which is easier to adapt.

At a simple like and see only.

So with the help of Open MP and with the help of the tools I can.

Engage the students to achieve the performance and how we can efficiently utilize the hardware in the CPU itself and achieve more performance and extract more performance from the CPU itself. So it will be easy for the coding also easy since we are not going down to the hardware level. So since we are using open MP.

For undergrad students, it will be easy to adapt uh, and learn easily. And even uh, I handled uh, uh, uh, I taught students M tech students with this open MP and parallel, how we can achieve performance with the experiments and examples I taught students already, I have experience, uh, in teaching students in VIT. In this open MP and parallel programming.

Introducing undergraduates to parallelism through OpenMP is an excellent way to connect them with practical performance optimization techniques. Finally, could you share an example of how you integrate structured teaching approaches to ensure clarity and comprehension among all students?

Yes, structured. So whenever we start teaching, so first we, we should focus on uh, conveying the concepts clearly, that is the fundamentals clearly to the students first. So once they are clear with the fundamentals, we can engage their understanding with the help of discussions, especially with the discussions, we can get the students, uh. Students listening skills in the class itself instead of teaching 1:00 to 1:00.

Through the uh, uh, through the type, uh, through the method of discussions, we can engage the students in, umm, understanding the strong fundamentals. So once they're comfortable with the fundamentals, we can move to the hands on applications. And my industrial, uh, experience helps me to engage them with the, uh, theoretical, with the real world scenarios. How we can map the theoretical thing with the real world applications. How?

3 The concepts are using. So this my industrial experience helps me to make make them understand more clear way in a more specific way. And for evaluating we can go for hands on examples and simulations experiments. Through those things we can make them make them understand more about how theoretical thing maps with the hardware as well as. In the application point of view.

Conveying clear fundamentals, fostering discussions, and connecting theoretical knowledge to real-world applications through hands-on examples is a commendable teaching approach. Let's delve into a few final areas of your expertise. Do you have experience conducting industry projects or consultancy?

Yes, actually I'm in uh, I'm working in the industry. So when I, uh, when I uh, start working at the VIT, so I'll be uh, happy to work with the industry persons because now itself I'm working in, in my organization, a professor, we are already collaborating in a research area. So if I come to vit, so I'm very much interested in uh, industry collaboration through in the contacts.

Which I have achieved with the various industries through my experience.

In the in the area of high performance computing and performance GPU performance optimization, I can since it is now industries are working in AI in performance optimization. So I can easily work with the industries in projects with the help of resources available in VIT and the students along with the students, I can work on the projects along in the industries.

It is valuable to have both industry connections and the intent to engage in collaborative projects for educational and research purposes. We will pause our discussion here for now. Professor Prakasam, is there anything further you would like to elaborate on or any questions you would like to ask?

Yes, actually, uh, now I'm working in an industry and uh, generative AI projects in NVIDIA and uh, Amazon, uh, project Amazon agent Ki project. So it gave me wide knowledge about how we can incorporate the high performance computing with the AI nowadays. So in, in, in NVIDIA, the LLM generated code, so we uh improvise the performance of the LLM generated code with the GPU.

Performance optimization techniques. So this gave a wonderful knowledge about how high performance computing can be incorporated with AI. And I was it, it aligned with my research interest. So this gave me a nice knowledge and I'm very, very much eagerly waiting to share with the students in VIT.

That is impressive, Professor Prakasam—working on generative AI and performance optimization at NVIDIA and Amazon aligns well with your expertise and aspirations. Sharing such cutting-edge knowledge with students will undoubtedly enrich their learning experience. Would you like to share anything else or should we conclude?

Uh, yes, actually, I'm very, as an, a proud alumni of VIP, I'm very much eager to work in VIT because I know I personally, uh, saw how uh, VIT is engaging faculties in the research and, as well as I'm very, I'm, I'm also passionate in learning and, as well as teaching as well as mentoring students and, uh, my research skills. So since VIT was offering multidisciplinary research and industrial.

Collaboration and collaboration with the foreign universities. So this is the very good place and correct place to work and grow along with the VIT mission and with the students. I'm very much happy to work with VIT. So this is what I want to share.

Understood, Professor Prakasam. Your passion for mentoring, research, and teaching, combined with your admiration for VIT's vision and opportunities, make your enthusiasm for this role clear. Thank you for sharing your experiences, expertise, and thoughts. Should we formally conclude the interview at this stage?

Yes.

No, uh, this is what, uh, if anything needs to be, uh, discussed, I'm ready for the discussion. So, uh, this is what, uh, these are the experiences I have gained in the, uh, in vit as an assistant professor, as a research associate and as well as an employee in employee in the industry. These, these are the experiences I want to share and with you, if is there anything to be.

To confirm, are you saying, "Yes, you can end the interview"?

Asked I can. I'm ready. Ready to share with you.

Thank you for sharing that, Professor Prakasam. You've provided comprehensive insights into your experiences and aspirations. Let me check if there are any additional topics to cover or questions to ask. Please hold on for a moment.

Please.

Professor, could you tell me about a foundational concept in your field that you particularly enjoy teaching, and why?

Yes actually, uh after my M tech I uh, I even I worked as a part time uh part time employee during my M tech there I taught students. I was in continuous connection with the students so and I enjoyed their I taught students the bachelor subjects with the students, B.Tech students, uh bachelor students, so they uh uh.

There I could see from the results I'm able to convey the concepts clearly.

To them and they make understand and, uh, by seeing the results and they by seeing their, uh, improvements, I'm very much happy. And, uh, I saw that my passion is along with the teaching. So after my Mtech also I joined TCS and during that time also I was, uh, taking classes within my, uh, alumni Mtech, my junior students I went to.

Periodically I went to take classes and in my vit also.

Close contact with the students as well during labs in projects.

Every time I'm associated with the students and I handled, I'm, I personally handled uh, classes for M tech students in parallel programming and high performance computing. And I gave lab sessions in open MP and UH, even my Mtech, we handled and workshop with a group of 20 members. I myself handled the session. So, uh, I'm very much interested in sharing the knowledge with the students and it is an. So this this became my passion even it.

My industrial experience.

Gave me more confidence that I can share the knowledge and make them.

I can make them fill the gap between the academia and the industry through my industrial knowledge. So I can share more more than the theoretical with the industrial experience. I can share more to the students and make and and help them to fit for the industry as well as in the research.

Uh.

Professor, I understand you have significant teaching experience in parallel programming and high-performance computing, coupled with close mentorship of students. To delve further, let's examine your research background. You’ve mentioned projects like developing a CUDA performance benchmarking framework and designing cache-efficient parallel Bloom filters. Could you describe how insights from these projects have informed your teaching approach, particularly in helping students connect theory to practical, cutting-edge applications?

Yes, actually in a performance optimization, if you come to the, uh, performance optimization, so, uh, we could see how, uh, data is, memory has been data is retrieved from the memory that is memory coalescing and divergent and parallelism. How we can explore parallelism loop and rolling techniques, those things, uh, in this project only I could see.

How we can improvise the performance? What are the ways? There are plenty of ways.

It's true, improvise the performance. So some of the techniques I come to know is warp level parallelism, reducing the warp, divergent branch prediction and loop unrolling and memory collaging techniques. These techniques I got from these projects I come to know from these projects. These are the techniques. So for Btech and Mtech students in their curriculum there they have a computer architecture and parallel programming in their curriculum.

So with the uh experience I have got from these projects I can give.

Theoretical as well as the hands on experience, uh, for them, uh, with the, uh, experience I have gained from the projects as well, projects which I have handled and, uh, during my research, what are the, uh, technologies have been going on with the research literature reviews, I have gone through the real time and nowadays, what are the research going on these techniques, I can give more insights to the students rather than the theoretical along with the practical.

A few for the bachelor students through the computer architecture, parallel processing and microprocessor, there uh, how we can access the hardware, uh, to achieve to get more performance out of the hardware.

Thank you, Professor. It's clear that your research on performance optimization has deeply enriched your teaching by bridging advanced concepts like warp-level parallelism and memory coalescing with practical applications. Let me ask this: when guiding students through such complex topics, particularly at the undergraduate level, how do you ensure their comprehension while also helping them build the foundational skills required to explore advanced concepts later?

Actually, for teaching these harder concepts, we should always, we should have a real world examples. For example when we are comparing with CPU and GPU. Instead, if we tell about CPU, GPU here, more processors here, uh, less processes, less cores. If we simply tell like that, say they will be, it will be hard to understand some. So we need to get some analogy and practical application for.

Example, if you go in a bike, we can umm 2 members can reach there. If we go in the bus 200 members can reach there. So in this way, we can compare CPU and GPU so so the throughput is less than CPU, the the throughput is more in GPU. So like this, by getting the analogy and real-world examples through comparing with the real world examples, these concepts, it will be easy for them uh, to uh, understand the concepts and uh.

May maybe it will be easy for them umm, to implement these concepts in the projects or forthcoming uh areas.

I see, Professor. Using relatable analogies like the bike and bus comparison to explain CPU vs. GPU architecture is an effective way to convey these concepts to students, especially those at the foundational level. This approach of connecting theoretical concepts to intuitive, real-world scenarios likely helps students build both understanding and confidence. Now, given your industrial experience and research in high-performance computing, how do you guide advanced-level students, such as those pursuing M.Tech, to explore real-world problem-solving in their projects? How do you prepare them to contribute directly to research or industry upon graduation?

The strengths and weakness of the of the research students while while analyzing a problem, what are the steps we need to since my since my research gave me these insects very clearly, so first I'll ask them the students to understand the problem first.

Yes, actually for for masters students first I'll ask them since I was doing research and I know where we.

So then from the problem first, then detailed literature review has to be done. So with these literature review and understanding the problem, they'll be able to find what are the.

What are the areas of improvements in a particular application or particular area? So based on that, they have to go for the experience and result analysis during which they have to document their progress. And I'll also encourage them to public for the publication as well as the pattern.

And which will be useful for their industry as well as their academia or research in In this way I can guide the students in a structured way. In which ways they have to start and they have to proceed and they have to with the frequent reviews, I can guide them in a structured way.

Understood, Professor. Your structured approach of guiding advanced students—starting from problem comprehension, conducting in-depth literature reviews, identifying improvement areas, and ultimately focusing on implementing and publishing their findings—demonstrates a strong commitment to both academic rigor and practical relevance. Encouraging publication and patenting adds significant value to their academic and career trajectory. With this insight, we’ve covered how you integrate your research and industrial experience into teaching. Let us proceed.

Professor, could you provide a detailed 3-year research roadmap that highlights both your academic goals and how your work will contribute to this institution's academic standing?

Yes, actually first, uh, three-year goals means first I'll make, uh, I'll, uh, teach students, uh, with the advanced concepts and the labs. And meanwhile I'll, since my research is in the way of CPU in multicore parallelism, I will in the industrial I in the industry I have gained in GPU. So I incorporate these techniques and how we can.

Uh, effectively use the highway for the Bluetooth application, the application we, we can use in GPU itself, how we can harness the hardware and achieve the more performance and it. Meanwhile, I can arrange guest lectures through guest lectures for the students with the industrial experts, as well as collaborating with the industrial persons for the research work. Since I have started already with the with my colleague, so I can continue.

So collaborating with them in the research and as well as the industrial consultancy projects with the high performance computing. So these are in line with the with the three years goals. So first step is for students teaching and mentoring and assisting students in the research projects in the high HPC. And in meanwhile I can arrange the guest lectures for the students with the industrial personals.

And. For the.

No. Umm, also collaborations. Industrial collaborations, yes.

Understood, Professor. Let's delve a bit deeper. How do you plan to mentor junior students in your lab to ensure a consistent pipeline of high-quality publications and maintain rigor in research output?

Yes, actually, uh, for high quality publications, uh, from day one, I, I, I will, uh, handle projects. I'll, if I'll take a problem and I'll split the problem into groups so each group of students can handle a module of work. So finally we can.

Combine the work and we can make, uh, make it as a research paper and communicate it to journals.

Uh, with the, with this point of view, students also gain knowledge in the practical scenario along with the theoretical perspective and the current research trend. They also aware, uh, how, uh, what are the works have been handled, what area of improvements can be done? In this way, research can be, uh, done with the students and the facilities available in VIT.

Thank you for that explanation. Lastly, could you elaborate on how you plan to utilize international collaborations to boost the impact and citation metrics of your research?

International collaborations, actually, my colleague, which, uh, who is working in Turing, is already working in, umm, high performance computing in GPU. So since we both are aligned in the same work, when we have a discussion, he know much about VIT and he is very happy to work along with us. Also, he is associated with IT university, uh, of Copenhagen in Denmark. So, uh, I'll, since we have already started collaborations.

Easy for me to collaborate with them, international with the universities and internationally in Denmark, and also I'm having. Contact My ex colleague is working in global foundries Germany, so in these ways I can collaborate easily with the.

Members available in abroad.

Understood, Professor. Let's delve into this idea of collaboration further. Could you outline a specific strategy or plan you would propose to establish and enhance international collaborations with universities and organizations abroad? Additionally, how might these collaborations benefit both the department and the students in measurable ways?

Time is up! Ending the interview now.

OK. Thank you.