Good day, Thileepan Stalin. 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?

Can you hear me?

Yes, I can hear you clearly. Let me know how you'd like to proceed.

Yes, I can hear you.

Proceed.

Professor Stalin, I'll begin by verifying your expertise and experience in computational modeling and associated skills. Let's start. Could you explain how you would design a computational model for a material undergoing additive manufacturing processes, ensuring accuracy and efficiency in simulating real-world behaviors?

OK, Thanks for the question. So to design A computational model for additively manufactured material, the first thing we need to understand is like what is the type of material it is and what are the process parameters used for printing these materials And the process parameters of these materials influence the final decide properties of the 3D printed path. So the best.

Understand what are the materials and what is the processing parameters and how does it.

Ideology is to.

Interact with the microstructural properties of the material and.

Define what are the mechanical properties affected by these microstructural changes. So understanding this micro scale FX with the process parameters helps us to define the structure process property.

So we can define the model mathematically.

Understood, you're focusing on capturing the relationship between the material properties, process parameters, and microstructural changes. Let's probe a bit further: how would you validate such a computational model to ensure its predictions align closely with experimental results?

A good question. So most of the models need to be validated. So to test the whether the model works perfectly, we need to do.

Experimental data with a few analysis on the experiments based on the type of property which we are interested in.

We are interested in mechanical property like tensile, compressive, 3 point bending. For four point bending we need to understand those factors and.

Consider the.

Other factors like what is the thermal stability, electrical conductivity and perform those kind of dynamic analysis as well. So by interpolating these data and collecting the experimental data, we can define what is the experimental results for this particular process parameters. And once we have this data, we can map these datas for the university inverse design and with this kind of inverse design model.

We can validate whether our predicted.

Properties and the experimental results match or coincide with each other. So this error percentage will allow us to say whether our model is 80% accurate, 95% accurate. So this is how we validate the model. It's similar to machine learning. So we have the predictive data. We should also experiment these things and find, finalize which is the correct value. So by fine tuning this, we can fine tune our computational.

Which we built upon.

I see. Using experimental results for mechanical and other relevant properties, combined with error quantification and inverse design, seems well thought out. Now, considering additive manufacturing processes may have inherent variabilities, how would you ensure your computational model remains robust across such uncertainties?

So that's a very good question. So basically we can use BCN based optimization where we can consider DOE and we conduct a certain type of experiments, capture the data of the experimental data, feed the input of the experimental data into the DOE to suggest the next set of parameters for performing the DOE. So by iterating these things within. 20 to 30 iterations.

We would be able to get a right parameters and fine tune computational model.

So it is similar to learning. Basically we are using the existing process parameters to evaluate the performance and use the performance into the data for training and predict the next set of variables such that we are capturing the exact process parameters how it influences the properties.

Thank you for providing that clarification. Let’s shift focus slightly. Can you explain how you would approach applying AI or machine learning techniques to improve computational modeling in materials science and manufacturing?

That's a very good question. So most of the knowledge which we captured is based on understanding of the real world manufacturing process and the materials information and this information.

About the materials and the process parameters are alphanumeric variables. So these alphanumeric variables.

Can be trained using the existing models. So LLM is a wide range of models. The basic thing is like with LLM.

A GPT kind of stuff. What we can try to understand is like we can have a collective knowledge of information with equations in build. We can understand how these things can be used instead of using an engineer based.

An engineer can design things based on the his previous knowledge and the books which he read right? Basically we can use a seamless technology like we can use the LLMS and GPT leverage on them, use their.

Knowledge and we can also provide the knowledge of the material process parameters and come up with a set of metrics for.

Defining what? What is the best parameters to print and what is the parameters which affects the influence the process properties a product properties?

So.

Sorry.

A for middle discovery is a new topic and it's upcoming and we can do this thing with self driving laboratories or autonomous labs. Basically we can perform certain tests instead of just collecting the data. We can also use this data to.

Enrich the model with the sufficient data to predict the new outcomes. So by these kind of approaches we can use AI for metal discovery and additive manufacturing.

I see, you aim to integrate autonomous experimentation with advanced machine learning to accelerate discovery and optimize processes. For practical application, how would you ensure the adaptability and scalability of such AI-driven frameworks across diverse manufacturing setups?

I would like to clarify on this particular aspect. What type of materials are we dealing with?

Good clarification. Let’s narrow it down to polymer-based materials commonly used in additive manufacturing processes. How would you ensure adaptability and scalability of your AI-driven frameworks specifically for such materials?

OK, let's thanks for the clarification. So for the polymers, there can be a different types of polymers that some polymers are.

Thermoplastics.

Also comes under polymers and another set of polymers are two-part polymers which on mixing starts to cure. So these two have a different types of extrusion methods. In the first method like a thermoplastic, let's assume the temperature of the nozzle and the printing speed and deposition rate defines how the metal is.

Such that the properties of the material and cooling rate of the material defines the structural property of the material at a microscale to the macro scale. So these exhibits a wide range of properties. And there is also a study which a lot of studies which says like the tool path or deposition method of the printing, printing along with the print process parameters place a crucial role in deciding the mechanical properties of the material.

Is deposited in a 3D space.

So basically this is a kind of a holistic approach We need to take on defining how we can better develop a computational model for this approach.

And if I understand correctly if it is.

Computational, uh, design for polymers, especially thermoplastics, it's.

Temperature of the nozzle.

Deposition rate.

These are the.

Crucial factors under tool path.

These are the three crucial.

Thank.

You're welcome. Let’s continue. Can you discuss your experience with teaching both theoretical concepts and conducting laboratory sessions, particularly in computational modeling or related subjects? How do you ensure an engaging and structured learning experience for your students?

Honey, your voice is breaking so.

If I understand correctly, you are asking about the teaching experience.

Hello.

OK.

Great. So I I worked as a assistant professor in a engineering college where I taught subjects like engineering mechanics, dynamics, kinematics of missionaries, and also engineering graphics and robotics for starting from undergrads from first year to final year students. And I also conducted laboratory sessions for mechatronics labs. So these were my prior experiences.

Apart from the during my PhD, I also did teaching assistants for different subjects ranging from engineering, design, innovation, healthcare, product development and for all these subjects the main thing our core is like math behind it and designing of the formulas and implementing the formulas to.

Arrive at a solution and if I.

Apart from this, I also attended few other computational courses.

Online courses connected by MIT computational modeling and also attended summer schools in Denmark on computational design and fabrication and all these knowledge accumulated helps us to gain a wide range of knowledge in this particular field from a diverse.

UH domains experts from different parts of the world and these are conducted by faculties from MIT and University of Copenhagen and these collective knowledge accumulated with my prior experience of teaching and my teaching in Singapore University of Technology and Design combinedly gives me an edge on advantage to teach the.

Latest and advanced computational modeling techniques.

Apart from this, I also studied.

Computational science as a subject.

So this gives me a platform to integrate computational based.

Approach a computational thinking into the process of building.

Process products and systems.

So I can handle a new curriculum on computational design, computational fabrication, computational modeling and computational science. Apart from that, I can also conduct laboratory sessions on these topics.

Thank you for sharing that comprehensive overview. Given your experience, how do you ensure that students not only grasp theoretical knowledge but also develop practical skills during lab sessions?

That's a great question, So.

I was an alumni of VIT Vellore as well. I was exposed to project based learning and research based learning and that is one of the advantage which I gained is like not just.

Putting these concepts which is learned in the class into projects and research so by that student can gain the knowledge. And this is carried forward in the SUDD where I carried my PhD as well. So SUTD follows MIT curriculum where we focus mostly on project based learning and theory. Subjects are like a 40% weightage and 60% is for a practical.

Attending the classes and memorizing the concepts apart from this.

Focus on the project development for the particular subject which enhances the skill and reinforces the knowledge which is gained through the theoretical classes.

Projects, so most of the sessions.

Learning and projects helps the students to gain a better understanding of the subject and a grasp on the subject and this is very helpful because by gaining the knowledge of that particular subject using the projects and research, they can be able to apply these skills in their real world problems and in their jobs as well.

So by this kind of practical research based.

And the most important thing is like there can be an option for students to convert these projects into research papers and can present in conferences that gives them a wide visibility and improve their CV. Apart from improving their CV, it also gives them a confidence that they are able to implement the concepts which is learned in the classroom.

Settings.

That's an excellent perspective, emphasizing both project-based learning and research integration to deepen understanding and real-world application. How do you balance guiding students closely during such projects while still encouraging independent critical thinking and innovation?

That's a good question. So there needs to be a a systematic evaluation. So we grade them periodically in phases like phase one, phase two, phase three and a final demo. And there should be also a peer review between the students to mark the contribution made by the students. And this is mostly a positive peer review which gives them.

The scoring by the peers and also the ways to improve their contribution to the project in the upcoming phases. So by having these systematic evolution across the phases even as a slow learner or a slow contributor, low contributor can also contribute in the upcoming phases by giving these kind of feedbacks to the students in the areas which they can improve on also.

This is very useful for people going forward with the industry setting where they can use this teamwork and they can understand their.

Helpful to make them work in IT a team and academic environment.

Drawbacks and make it into a positive approach of learning.

So systematic evaluation and peer uh.

Peer grading will help us to evaluate students on a regular basis and give them a proper input for them to improve and improvise and grow.

Understood. Your approach of phased evaluation and peer feedback fosters both accountability and collaboration effectively. Let’s shift focus. Can you describe your approach to guiding student research and ensuring their projects contribute meaningfully to the broader field of computational modeling or materials science?

Sorry, uh, the voice is breaking. Uh, can you repeat the question again?

Certainly. How do you guide students in their research to ensure their projects contribute meaningfully to the broader field of computational modeling or materials science?

OK. That's a great question. So.

As a contribution, A contribution comes from multiple perspectives. It's a contribution to a new knowledge creation, contribution to the society, contribution to the industry or economic value. So these are different types of contributions which a student can make, a project can make with the research, right? And then students learn computational modeling, the way in which they approach the problem.

Solve the problem and implement the computational modeling into the system.

Problems which they are facing is a given approach to solve a better solution. So this is needed because students need to understand the value of their project or the project or the solution which they developed for a particular problem to evaluate that they should also be thought about the metrics of the impact.

The impact is what makes it impact in terms of number of people.

Lives. It has improved. The solution has improved.

And the impact of the economic benefit which?

The industry can gain with this particular solution of the computational modeling. So these kind of values and metrics to evaluate their solution.

And the project is necessary to evaluate or unders to make the students understand the impact they can make with their project on computational modeling.

That's a.

Thank you for sharing that. Let’s explore further. Can you elaborate on your experience with publishing research in reputed journals? How do you determine impactful topics, and what has been your strategy for ensuring acceptance of your work?

That's a great question. So when we do scientific research, the main question is like what is the problem we are solving? How are we solving it? Why are we solving it? By asking these three fundamental questions, how, what and why, we can understand the audience and the sector and where our solution is going to be impactful. And we made a few papers, I would say.

Uh, not few papers more than.

20 plus papers in this research field which are widely presented in reported international conferences like ICRA, Robosoft and other garden research conferences. Apart from these we also publish in a high impact factor journals where we got a very good reviews and comments and feedbacks for the papers. So technically. Paper Writing.

Is a skill which can be acquired by uh people.

Who undergo the training of doctoral program or a PhD. So I underwent APHD in SCTDSCTD is a Singapore Ministry of Technology and Design which follows a curriculum of MIT and I was guided under professor who completed.

Bachelor's, Master's and PhD in MIT. So he transferred whatever knowledge he gained over his period of time in MIT and I followed his footsteps and.

Apart from publishing papers, he also insisted on publishing high impact papers instead of just focusing on the number of papers published in a year. So instead of being a paper mill, we focused on papers which really contribute to the scientific knowledge and the impact which it can provide to the society.

Apart from this, we also, uh, patented few technologies which we believe has economic value and a benefit for the society and the industry.

Thank you.

I didn't underst, I didn't get it. So there was a pass. Did you say something?

Professor, to begin with, can you

Sorry.

Yes, Professor. To start, could you share which introductory computational modeling course or topic you find most impactful and rewarding to teach, and why?

I would say computational thinking as a first introductory course because competition comes from a base of thinking how we can solve or approach a problem. So basic introduction to a computational thinking is the first fundamental step to be taught and mathematical approach of solving a problem for.

Boundary conditions has to be established because all the problems can be solved if it is well defined. So to reach that particular point to define a problem, a person needs to understand what type of problem it is and how to approach the problem. So the introduction to computational thinking is a necessary subject a student can gain to get deeper into computational modeling.

A particular.

Thank you for elaborating, Professor. Let us proceed.

OK, so.

Sorry.

Go ahead.

OK.

Specifically, describe the research objectives, methodologies you implemented, and the academic or practical impact of the project.

To start with 3D printing robots, soft robots. Soft robots are made with soft elastomers textiles. And if you combine these textiles and elastomers you can make fiber composite structures. So fiber composite structures was a initial idea of making a better robots so. Building on this building block of making fiber composite structures.

Into soft elastomers and these soft elastomers are super stretchy, flexible and it undergoes 1000% elongation at break.

With the differential hardness and the one way to improve the mechanical properties is by embedding the fiber into this silicon matrix to do that.

We need to have a repeatable, scalable approach is to use a 3D printer and there is no printers in the world which can do embedding of fibers into silicon matrix. That's where we.

Developed few processes and techniques to embed long continuous fibers into silicon matrix and on the exploration we found that.

With functional properties. So to tune the functional properties we can embed continuous fibers which are conductive in nature such that it can work as a sensing tracks.

When we the mechanical puppies can be tuned along with.

Conductive tracks can be used for transfer of data, signals, power. All these enables us to create a wide range of objects and we were able to achieve different types of geometries and structures by embedding these kind of functional fibers into silicon matrix.

So the goal is to make 3D printed soft robots. As I mentioned before, a robot is made of sensors, actuators, control and communication. So all these 4 components are very important. So we've approached this with a very sensitive.

Balanced approach with the incremental research on developing sensing coils.

Wireless power transfer to charge the batteries. We also developed NFC coils to communicate with the nearby robots. We also developed antennas using this.

Embedded coils to communicate the data collected. So it's a very balanced approach of developing sensors, actuators, control and communication. So we developed this automated fiber embedding as an approach to develop 3D printed software and we have a a patent granted for automated fiber embedding in Singapore.

Thank you for the detailed explanation, Professor. It appears this work blends advanced materials research with practical engineering capabilities. Could you elaborate on how you would translate such a technically advanced concept into classroom instruction? Specifically, how would you introduce foundational ideas from this research, such as fiber embedding or soft robotics, to undergraduate students with limited exposure to additive manufacturing?

That's a great question. So these knowledge is accumulated is based on the research as a fresh student who is exposed to these kind of techniques or technologies is a.

It's a bit hard to grasp these technologies in a classroom setting. The easy way to expose students to this kind of.

We'll have a theory class explaining how this works, how the concepts work, what is the principle behind it, and what are the parameters involved and students will be exposed to.

Advanced Research is by practical experience, so the students.

Have a hands on experience to develop or use this automated fiber embedding techniques such that they can understand the pros and cons of the process.

By that approach, students can also learn to implement the technology developed.

For research in a classroom setting so they can also implement these some of the learnings into their projects as well.

So this kind of approach of lab based.