Unleash The Full Power Of Data: IISc | TalentSprint's Computational Data Science Programme
Good evening, Professor beeper. Good evening, Professor Shashi. Really good evening. I am Great. Thank you so much for joining us today.
I believe this is the first webinar that we are doing for the computational data science program that has had a very successful launch with ISC. And it's a pleasure to have both of you here today. So, we will get started right away, we have an hour long agenda ahead of us, the professor's have a presentation that they are going to be sharing. And, and we are then going to be opening up the session for all of your questions. Firstly, thank you for joining in today. It's been lovely having you. You can send your questions that you have I see a couple of people that send in their questions, we will do a dedicated q&a section at the end of the session, around 15 or 20 odd minutes. So I'll request you to hold on to that. But you can send in your questions in chat as well. I hope the screen is visible and you're able to hear us all clearly. Yes, on the chat window will be appreciated. Okay, thank you, Prashant Jain, thank you so much for the confirmation. So, we're going to get started right away. So like I mentioned, you know, my My name is Ernesto Bhattacharya, I am a senior director at talentsprint, I primarily had all admissions for negative programs that we run in the company, one of the programs being computational data science. When we launched this program, sometime late last year, we were, we were not anticipating the kind of response that the market would give to the program, we initially thought that we're going to get started with just about, you know, 50 odd people in the cohort. But when the classes started on ninth of January of this year, we had to double the class size, and it started off with about 100 101 people persons in the class. So it's been a phenomenal response that's there. And, you know, before we get kind of started into the program and answer a lot of your questions, I have some of the questions that, you know, we've been receiving, from some of our panelists from some of our participants, or people who are considering the program, we're going to get to that. So just starting off with something about the current patch that we have in class, the that the professors are teaching, it's a great diverse batch that we have, and it's it's lovely to see, you know, 20% of the class as women, the diversity in the class in terms of years of experience is around, you know, you have somebody who has just about graduated, too, Europe, somebody with about 32 odd years of experience, on an average, we have about 12 plus years experience, people coming in Indian cities, international locations are all there, there are more than 18 sectors represented 87 companies in class, which means that the amount of knowledge share and that's something that's very important in in these kinds of programs, because a lot of the learning happens outside of the class and it happens you know, in in the groups from the peers that you have. So, the larger the representation in class, the higher the amount of learning that you can get, because you can kind of pick up from experiences that people have in companies in sectors etc. So that means it was a wonderful he represented class. You know, if you really look at it, a majority of the people were engineers again, but you still have people who are joining in from you know, something areas like botany, economics, bioengineering, agricultural science, etc. And you had, you know, software engineers to directors, CEOs, vice presidents joining in so vast gamut of experience was there in the class. Some interesting people we are we had about 50% of people joining in from Bengaluru, but you know, it still the rest half of the class was not from Bengaluru. You know, Major, there's somebody who worked in a major sports league in the US, we had representation from the defence forces in the Indian Army. We had a couple who were there in class. And then there was somebody who had, you know, won a tech geek award in 2018. So this was, you know, this has been the journey of the current code. It's been a very small journey, so to speak, in the longer scheme of things the cohort has just started about a couple of months back, but it was a journey which started off very well. I mean, I you know, move to Professor Deepak and Professor Shashi you know, what do you think? expecting this kind of response when we had decided to launch the program?
Yes, so definitely, the data science field is growing. And the product that we are offering kind of the course that we are offering is been specifically designed for the industry professionals, and data science, practicing data scientists. And we were very confident about going with talentsprint and that you will be able to convince convinced the students also Yes, we did expect that we will be off to a flying start. And that is what has happened. So it is within within expectation, I would say.
So, what we are going to do is that we will move quickly over to, you know, the D Section A couple of things. One is that, I will request both of you to introduce yourself because, you know, that's something that speaks volumes dv could be the the faculty who are there you are one of the biggest asset probably in the program. And so I request both of you to start with an introduction for yourself, and then you know, we can move on to the slides that you have, and then we'll take up the q&a. Is that okay? Yes. You can probably start.
Okay, so, answer sheet more, I welcome you all, once again, for this session. I'm currently the chair of the department of computational neurosciences arise in the program coordinator for this advanced ground relational data science. So my expertise is on more quickly computing. I started my career as a mathematician, it's more of an applied mathematicians will do HPC on the data science. So I will be presenting briefly about all the faculty members, those who are involved in this program and what their backgrounds are. So apart from that, yeah, so we are having plenty of experience, especially running this impact programs on donation data science. And in fact, I see is the first Institute to start up and pledged department on computational data science that was way back in 2015. Because the full fledged department for computation, data science, but now you name it, almost all government Institute's as well as the CPAs and even the private Institute's right even I'm reading emails from the other private Institute's they also are started offering different variants of competition data science programs. So that's how the idea again, so we are running like PhD programs versus programs for data science on the impact programs. But these two things are really deadly. So that's the reason idea conflict, how to start with how to calculate the working professionals, I will be talking only about more about these details of this connection. So if this is stopped here, maybe after defects, then I will go over the details of this course. And yeah, thanks. So I am Deepak subramania. I'm Assistant Professor in the Department of computational and data sciences at the Indian Institute of Science. I have a PhD in computational engineering and mechanical engineering from the Massachusetts Institute of Technology MIT in the US, I also have a btech from IIT Madras. Mechanical Engineering, again, specialization in energy technology. I my area of research is on Bayesian learning and applications of machine learning and artificial intelligence with geosciences, remote sensing, climate change, all those kind of things. I also do significant bit of work in autonomous planning of vehicles, autonomous underwater vehicles, drones, and these kinds of things. And yeah, that's a brief introduction about me and the process. She talked about the other faculty were there and about the CS department. as a whole.
Yeah, thanks all of you. Okay, so I know how do you want to go about it? We are you we want us we can we can talk
with you. Yeah. So the idea is, maybe you can cover it up the major production data thing, then I will start and then I will give a brief overview about the how we arrive at this program, and then the faculty involved in this carbons program. Sure. And then we'll take over the project.
So what we can do is professor, we can probably talk a little bit about the genesis of the program, so to speak, and, you know, I can talk I can, you know, probably name some of the faculty and you'll probably be able to give A couple of Purell blinds about 10 Yeah, sure. So maybe I also did the presentation, I can also share so that people are they want to vote yes, you can probably share the screen lesser.
Okay, so I'm able to see my cspr. Okay, so just to like, I'm not going to take much time, because the more of this, we need to discuss the whole program and other things. So I will directly go into this program. So um, so how we started this course discussion, some other things are posted Shantanu on back channel, from the talentsprint hacker has through the team to start a computational neuroscience program. And at the time, when we came up with this idea of starting programs, right, of course, as I mentioned earlier, we have a good experience in running the research programs on the impact research programs and programs for two years. So we have good experience with But for us, the starting the two years program from the two year program to a shorter version, and depth to like the working professionals, is completely new idea. So of course, we have to customize it. And we have to come up with the new ideas what we are going to gain on being white in this program all these questions. So immediately, right, as usual, even if you want to start up by this while we want to know so what's the status of your field? What is the state of the art and other things? Right? We just immediately said, Why do you want to offer a data science course. And then don't go into the details, but you can really see the number of kids that you're going to get, but then of course, this is not going to give you any conclusion, but one point that you can infer from this place this page is of course, you can really see the how popular the course is find that it is this case on what we are buying, we have money online. So we are refining our searches, but still the data that we are getting are the hits number of bits that you are getting is not going to help us then the version counts like maybe in India we are offering it but though this is not the case you can see from the list of participants what it is it is not presently from India, but at least we are starting from India, we are searching in this second this way, right? Still, you can really see that number of bits. So obviously the question comes, if there are already plenty of content, materials, online courses, everything is available, then why do we need to offer this program and to whom you are on this ground? What is the unique feature of this product? These are the three questions that we want to address. And we have done a lot of homework behind this. That's the reason why the back has mentioned right? Why we are not really surprised. But of course, we have expected this going to go a big hit. Of course, it's exceeded our expectations. Somehow we have to stop our admissions that Okay, go with this 400 we do not want to take more. Right. So that's how the correct way we have started. The reason behind is we felt that these are the three questions that we have under us, maybe that is the reason why we were able to have funstar. So then the question comes to why because what are the new things we are going to bring in so then we are targeting the audience, right? Especially we are catering to the working professionals, and also the professionals, those are not having much time, they have to work in between or they have to do this program in between this their work? And also, what's the unique feature what there are plenty of data science courses, what are the things that we are bringing newly especially when you add the computational component to the data science.
So that's how we have started the discussion and after plenty of discussions and several rounds of discussions within ICT across every sprint, and I see. So we have come up with the modules or different set of modules. So we want to have our participants often successfully completing this program. We just don't want like a users of data science tools. But instead what we really need is we want our users our participants to really know what is going behind the tools. And they also should know like or reasoning, right? Especially in the data scientists have kind of gray area when you go to do some classifications are even doing that neural networking and even do some whatever the data science model that they are using machine learning artificial intelligence, everything is kind of an approximation and that one needs to know whether that approximation is intuitively correct and as well as the algorithm quickly on that what's the math behind it without knowing the maths where the reasons algorithms, we cannot really justify the the outcome because the computer will always provide some outcome. Right? So these are the reasons why we practice our program with these modules. For example, when you start with this module is competition data science in practice. That's I would say that is completely different from what you get from the existing courses computational what I really mean is we really start with how the computer works. It's all about the computational thinking we are bringing into the data science. So starting from how the numbers are represented. He say that okay, the whatever you're getting is having some approximation, and then how to use a viewer system efficiently. So in this goal, this module covers about the how to use the your popcorn machine, or our coordinations, even in the last year live in the production side, large scale problems, if you want to deploy it in your organization, then how to use it for large scale missions, cloud computing. So these are the ideas that we want to bring it down just like a data science tool. That's how the computational data science comes into the picture. And also the mathematics and for data center of neuroscience, it's all about the probability statistics, optimization, these are the basic tools, mathematical tools want to know, to understand why you are getting such type of results. That's how this that's why this mathematics of data science comes into the picture. And also about the data engineering, that engineering is it's not only about learning how the tools, how to handle the big data, how to deploy it from the cloud. So all these things, topics will be covered on this topic. And of course, talk about the machine learning, then it must be part of the data science, that's a whole, I would say, and the neural networks is the, again, the core of data science. So these are the tools under the core, it's not just you're going to use TensorFlow, or you're going to use like a Google colab. It's not like that. But you will understand what's going behind those tools. That's the difference, compared to the regular data science course that you're bringing in. Of course, this is mainly for the business professionals, then we also want to have like a module on how this can be deployed on how to think like a business person or business organization. That's why the Business Analytics code comes into the picture and how to measure it with this data science. So that's how the fit modules basically. And in addition to that module zero or module, the bridge course, which we know that we have heterogeneous group, and people might have left the educational institutes like 10 years, 15 years back, so they want to come back to the educational center. So that's how the course modules is done, at the end of the day, so it's not about only about the learning, it's also about how to deploy, right. That's why we come up with more than 4025 Mini projects. And also you will be having a capstone project that capstone projects idea is you bring with your own project, or one problem set of problems, and then deploy whatever you have learned out of it. This goes on the deployment, really solving the industries. And this will help you to build your career, we believe, and this also makes you like more hands on. So that's the idea how this is how the course structure is built up. So from there, then we have looked at the Faculty profiles, and when we talk about the faculty profile, so it's it's not about like, we want like two three people, we want to offer it or different persons from the same department. It's not like that. So we are then after fixing the course modules, then we started to look at the faculty across the Institute. And that's how we want to have the best faculty to teach the respective modules in the complex of data science. That's how we come up with the pool of faculty members, not only from Department of computational neuroscience from Eazy, e from management theories, and so on. So starting with debugging he has already introduced in his main research areas are based on learning machine learning, artificial intelligence, part environmental for past autonomy. And Professor Yogesh, human VMware is mainly working on the systems on the data computing systems, big data, and Internet of Things. And also he's going to talk about this data engineering and so on.
And then we also have professor at andare. And he is a kind of CLP person in the sense, but he's also doing a lot of great data driven science on he is also working more of an HPC than he is going to mainly work on are going to talk about this mathematics for data science. That's the core area which he is going to bring in. And then Saturday, when he comes from the Management Studies, he's going to talk about the cons as opposed to the research areas or quantitative finance, derivative pricing and the real option analysis, but he's going to talk about how this data science is used in business and how one can do the bridge bridging between the business and the data science and funding. Again, his areas of research are a combination of signal processing, gnn, machine learning, statistical inference, and so on. So again, he is going to bring the ISA Core Data Science person, and he works mainly on the neural networks. And you can really see and see from his research expertise. And he's going to teach about this module on the neural networks and for sacani marmari. He works on most of the things that hold them in machine learning for material design, and computational materials physics and linear algebra is again going to talk on this optimization, the statistics and so on. And in addition to that, as I mentioned, Earlier, right, so it's also we will be having like a bridge course under this bridge, the course is about to bring the, to this level of the boat or to start with the module one with the different heterogeneous groups. And this will be handled by Ahsoka and his team. And also going nice again, like these in the field, I'd say is one of the well known experts in Python. So he is working with Python for several years, probably, maybe the till the time the Python was invented. So he is having plenty of experience in offering this type of courses. He's also like a very big supporter of open source. And then he is also involved in this type of projects from IIT, Bombay. And those myself I have already produced, right, my area of expertise is mainly about numerical analysis and data science, scalable algorithms, I will be handling most of the module, module one about the computational thinking and data science. So with this, I just want to pass here, and maybe the deeper can continue or here any questions, we can also start, maybe I will postpone the q&a till the defect finishes, then we will start the q&a session. Is that okay? Or retro? How do you want to handle it? Yes, perfect. So
people can speak, we can have the q&a question. Some people are sending in their questions, you can send it on chat, we're going to try and answer all of
your questions. And then I can do the chat, replying to the chat. But yes, again, we can open up public.
We can see your screen capacity. But yeah. Great. So with that great introduction, let's look at what we mean really by computational data science, right. So that's, I think that's why you are all here to understand what this course is all about. What is data science? Right? So that's what we are going to see now. So if you see the word data scientists or data scientists, it's supposed to be the sexiest job of the 21st century, bringing in a lot of salaries, right? So the amount of salaries if you see, it's probably one of the highest in the industry, right for data since. So people often compare data to oil, right? So data is a new wine, they say which it has to be processed. And information has to be gleaned from the data by applying mathematics by using computers. Right. So that's where this whole idea of computational data science comes in. Right. So as Patricia, she also mentioned right, when you Google for data science, there are like these lots and lots of ads that come about learning data science, and more. What Wikipedia says is neuroscience is an interdisciplinary field that uses specific methods, processes, algorithms, or mathematics to extract knowledge and insights from data, right. So that's what data sts means. And similarly, you can kind of use Google search as a proxy for the popularity of certain terms. And if you do that, you will see that the field of study he does science and machine learning. So in blue, what you see is data science, and in red that you see over here is what is machine learning. So by about, say, 2012, as the there has been an uptick in both the use of data science and machine learning. So this has been normalized with the maximum search volume, right? So if you add in the term Python to the mix, right, so to the mix of data science in Python, you can see that Python has already been more popular than the word data seems. But corresponding to about 2012, when the data science starts to pick up, right, so about data center, typical Python has really picked up right, so really much more, and the steep growth in the use of both Python and data science. And when people usually ask the question, right, so what is artificial intelligence or machine learning better, which is machine learning, same as AI? Like what the hell is the difference between data science, machine learning, artificial intelligence, deep learning neural networks, right? So all these terms are out there and then people are really confused. So what what is really happening and part of the purpose of our next 10 to 15 minutes is to demystify this whole notion of data science and where the field lies so that you can make the most appropriate size for you in either enrolling for our program or what what is most appropriate for your resume. That's the goal for us. And usually, if you're asked to slightly the previous generation, And of industry professionals or even researchers, they might tell you that AI is a black box, right? So neural network is a black box. Right? So is that true anymore? Right? So has there been some understanding that neural networks are somewhat like why do neural networks work? Right? So all these things become really, really important. And then I guess in most of your minds, this question is there, right? So today, should I really study data science, or study machine learning? Can a data scientist become an AI engineer? Alright, so what is AI engineering? What is data science is A is a subset of data science. So all of these questions must be there in your mind. And that's what we are trying to address here. And the main issue, right, which I see is that we are in the jargon universe, right? So people use machine learning AI, deep learning interchangeably in conversation, and that trips us up, right. So we have drowned in the jargon us. This is a fact. The fact is that there are several tools and techniques that a data scientist must know when it is expected to know several tools or techniques.
But the main issue is that it is really hard to separate tools from techniques and that causes confusion of people. So let's try to see what is the difference between tools, techniques and methods. So in this context, so the only thing that you must remember is that change is the only constant, right? So tools change very, very fast. Right? So c++, Python, TensorFlow, Python, spacebar, Julia. So when I was doing PhD, Julia was not even there. Right? So in 2012 2013, that's when Julia started picking up at that time, if you said, right, so in around 2007 2008, when we started, at that time, if you said you are using Python people like okay, what is freezing? I don't know, right? So and people are using MATLAB and things like that. But within a span of five to 10 years, the tools are changed, right? So TensorFlow used to be really difficult to use about three years ago, you might have heard about the students are slow. But today, it's actually really, really easy to use, right? So people to TensorFlow 2.0 was released about maybe about two years ago. And things change very fast in the tools domain, right. So there are lots and lots of tools that come up, and they change extremely fast. But techniques, on the other hand, they change at a much slower pace, right. So regression classification. So these are techniques for machine learning techniques, right. So for data science techniques, these techniques change at a slightly slower pace. So it took about 10 years for the support vector machine. So in the 90s, support vector machine was the most important technique that people were using for performing classification. And many of the artificial intelligence tasks support vector machine. And for about 10 years, nobody contested that support vector machine was the king of the field, then it took another four or five more years for neural networks to merely imagine from about 2006 is when this use of deep neural networks have started become more and more popular, and in about about 10 years, alright, 10 or 15 years, neural networks are matured, and that does become the core, right. So it takes about 10 to 15 years, that's the timescale in which techniques that are popular for a decade or so change. But behind all this is the fundamental concepts, right? So that almost always need stays a near constant rate. So statistics, carpeting, optimization, matrix algebra, all these are the fundamental the base, the foundation of your building, right, so the foundation of your building stays constant for a much longer time. So you can think of these concepts or techniques or the fundamental ideas as the foundation of your building the pillars, etc. Right? So the floors can be the different techniques that you build on top of that, and the interior, which you can change, just like that, right. So that that would be like the tools like so don't get confused with people saying, Okay, this particular tool, do you know, this particular tool Do you know that particular to if you know, the fundamental concepts and techniques, and tools are really, really easy to pick up? So once the fundamental concepts are in place, techniques, tools are easily picked up, right? So this is the core of the learning process. If you really want to learn data science, machine learning, etc, then you must build from the foundation, then start building the flow state. So that's that's the analogy that we'll be using.
So let's try to understand so there's a lot of people who asked what is AI? What is ml, etc. Right? So let's try to frame that in context. Where does AI machine learning like, where does data science like right so let's consider Data Science University. So the universe of data science lies Yeah, let's let's say this green blob is universal data sets within that universe of data science. The first and most important thing is problem formulation like what do you want? Like, what do you want to do? Right? What do you want to predict the stock price for next week? Do we want to predict if we run corys, LBW or not? Right? So do you want to know if you have cancer or not? Right? So this is the problem that you want to solve. And so that is a problem that you want to try to address. within this context. There are several tools and techniques that you will use. And one of that is machine learning. The other is data engineering. The other is Internet of Things. Not everything is smart, no radio or TV smart. You're smart, you're for smart, you're watching smart, right? So everything is smart. So they all wait what what is smart, they are all internet connected devices that send and receive data. So as what is called as Internet of Things, then you might draw some inferences from using the data, applying your machine learning, etc. And then you want to communicate that to your stakeholders to your clients. Right? So then you have to say stories in terms of data, right? So data stories are really important, then you want to show nice visualizations of the data. So sometimes visualizations are very powerful, right? So you want to do that. And that becomes the data visualization aspect, right? So this is, within the whole universe of data science. All all these are different different elements that exist. Now, where does neural network come in? So neural network is one specific way of doing machine learning based on machine learning. And neural network is one way of doing machine learning, it's comes inside machine learning. Then, all this mix comes big data. So big data is really about having lots and lots of data. Now this big data intersects all these other aspects of the data science relevance. And here you see their big data intersects with machine learning, and specifically where it intersects with neural networks. That is that small piece of this entire real estate space, right, so that small piece is where your deep learning lies. So deep learning is the application of deep neural networks, specifically using lots and lots of data using big data. So it's a machine learning technique that uses lots and lots of data in a framework called neural network to solve problems, right. So that is what deep learning is all about. Neural networks and deep learning is that is where it lies. And where does AI lie in all this? Right? Well, what is artificial intelligence and all those which are data science universe has been defined? And in fact, artificial intelligence can exist with or without data? Right? So if you just write the program that says if else condition, it's an if else condition, if you do that is also an artificially intelligent program, right? So it knows what to do, if the button does not need any kind of data, train that. So it's a rule based system, right? So artificial intelligence can in fact, exist outside the realm of data science. And usually when AI intersects with data science, that portion is what is called as the machine learning most often. So and then a neural network is one such technique that in recent years, I've given a big boost to artificial intelligence, and then deep learning, per se, right? So within that, when you use a lot of lots and lots of data to learn all these things, that's where this whole field writes, right? So data science, artificial intelligence, machine learning, neural network and deep learning, right. So, this is where it lies. And when we say computational data science, right, so data science is an umbrella term, data science includes everything right. So each and every part of this green circle is included in data science,
traditional data analytics, so if you are from the data analytics field, right, for traditional data analytics, this is related to modern data science, and you can consider it almost equal in meaning, but with some several certain differences, and those differences mainly come arise in the use of high performance computing devices, like graphical processing units, using these high performance computing units to actually do computational data science. So if you perform all of this using high performance computing equipments, that is what is called as a computational data skills, you can in fact, do data science on Excel, like on your piece of notebook, right? So if you don't draw a linear regression line that is you doing data science, you're collecting data from a small experiment that you conduct in your lab, Physics Lab, or you're you you observe that particular day's bust comes at 3pm 4pm. Right? So you observe that and then in your head, so you construct a predictive model of when the power time of the bus will be there, right. So all that history, data science, but you really don't need to use a computer to do that. So when when do you need to use computer This is when the scale becomes really big, right? So when you have lots and lots of data when you have lots and lots of questions that you want to answer, and you want to answer this in a scalable and fast manner, that is when computers come. And if you do all these aspects of data science using this modern computing devices, that becomes computational data science, and that's what we are really teaching over here. Right? So that's what this course is all about.
Now with that, another question that people ask us, why is data science important now? Right, so whereas data science, all GCSE statistics, the same as data science, is mathematics the same as data science. So why data sets become becomes important these recently right over the last 10 years is there has been a sudden explosion of data, I said, everything is smart, right? Everything is internet connected device. Everybody, almost every every family has more than five, six mobile devices, which every person has at least two devices, two SIM cards, right? So these days, right? So these mobile devices generate a lot of data, there are cameras everywhere, CCTV everywhere, right? So there is traffic cameras everywhere. So all these cameras collect abundant amount of data, and the digital footprint of customers, right? So Google, Facebook, all these internet companies, they collect tons and tons of data, and that that explosion of data has made it sure that you need to use computational high performance computing and all these major infrastructure to actually use the data and draw insights and that's where competition data science and data science become really really important. in this field, you can see the amount of data that is being generated in just one minute to 2019 right. So, one average meant in 2019 has generated so much of data right. So, that is the amount of data that is available, that is what is making data science really, really important today. So, I will just give three quick examples on how specifically data science comes into practice right. So in finance, right. So banks, so you generally make predictive models to do real time simulation of market events. So, that is one thing that you do fraud detection, right. So, if you have a credit card, you swipe it in particular supermarket or something like that, and immediately the transaction has to happen, right. So, at that time, the credit card company has to decide if you are the one who's actually swiping it or there has been some fraud. So, if these days, you will immediately get a message saying that okay, this particular transaction that you did is not typical of you, right. So, there's a fraud detection thing that happens that is a particular machine learning techniques that they use to actually do fraud detection. So, you need machine learning in that way, that's how the finance industry uses it. The other major thing is in lending right. So, if you apply for a loan, and how they decide the EMI is equated monthly installments that you need to pay, right. So, in many many such cases, your data about what you do, what your age is, where you live, what is the kind of salary that you draw this data is put into a complex machine learning model that will then make the credit card company in more if you are trustworthy or not, what is the interest rate I should give you or not right? So all those decisions are also taken data driven these days using machine learning using my tools like neural networks using all those tools, right. So that is one example. Another example is that you know, and which many of us are day to day VC is Hawkeye decision review system, right? So the decision review system that is used in cricket. So what it does is is basically a predictive frame prediction problem. So video prediction problems. So they use high speed cameras to capture each and every moment of that ball and see how it will once it hits the pad, how it will progress from there, right. So you need to actually make that prediction. And then today according to today's capabilities, only half of the bar right. So, the precision is just about half of the bar that is why in many cases it goes to umpires decision, right? So that is because of the limitation in the camera imaging and the algorithms that are used to actually make those predictions right so that we are done using machine learning conville convolution neural networks and predictive networks. So that's what is used. And here you can see messy wearing some kind of vest. So this actually collects data about how messy is running and then the courts can make a decision about whether he should be called off and substitutions should be made or not. Right. So his points weighed, all these things it mentioned And then it goes into predictive models that classify whether he is fatigued or not right. So, that classification is done. So, all these things are there in sports. And in another example that I will talk about this is fresh from the research, right. So, I was just working on this today yesterday, this is the latest of the research field, where we are actually using an autonomous underwater vehicle that has to go out the test to find out if there is an enemy vessel that is present. So, this is autonomous vehicle that comes out it uses sonar, radar, vision cameras, it has different sensors that are there, and it has to actually see if enemy missiles person, it has to go and observe what is the temperature in an ecologically sensitive area, all these things, these autonomous vehicles do by itself underwater, right. So now it requires vision, computer vision it requires, and it also needs reinforcement learning machine learning algorithm. So this is how you model this problem as a machine learning, reinforcement learning problem, and then you try to find the optimal policy from it. Right. So, if you are interested, we have one paper published in that one more is being submitted. So this is just latest research from our group in ISC using machine learning for several applications that you may not be aware of the finance and cooking, maybe everybody knows, knows that.
So the final thing that we will discuss is why should you become data scientist, right? So this boom has been there for five to eight years in the past, right? So four to eight years, people have been talking about data science. But you might be surprised to know that in 2020, in a major survey that was conducted in the fortune 1000 companies, and the theme of the survey was data driven business transformation, connecting data investments, which investment these companies are made to business outcomes, are they connected, they this is what they were trying to do. And you might be surprised to find that in most of these big companies really investment in big data, and AI has kind of leveled off, right, so almost 99% of the companies said they have already invested Wait, so they're not increasing their investment. Right. So it's leveling off. But only a very small fraction has actually put AI in production and use AI and data science in their business process. So people are trying to find out why right, so we need to answer the question why in order to progress. And what most of them said is that they don't have a data culture. Right. So only 27% of the companies actually say that they have successfully built a data culture so that what we're doing is people and business processes, that there are no data scientists available the people who are working on the domain don't know about data science, how to use the data to leverage the data, they they are not aware of that, in most of the cases, technology was not a challenge. The tools were not a challenge tools, technology, everything was available, it is people who knew how to use data science that was what was missing. So, this was from a major survey. And this is in fact in companies This is what is called the viability gap right. So from theory to practice with So, the gap is usually this is called the innovation valley of death. So, people know people are invested in data infrastructure people have invested in companies are interested in data infrastructure, but to go from proof of concept to a minimum viable product. So that portion is what is called as the gap innovation when you have that and to bridge this innovation valley of death, we need data scientists, which we need people to champion people who understand how to build data products and pay it out and that is exactly where our cost comes right. So, this is the need for data centers, new data scientists with knowledge of all the concepts, techniques, tools, and domain domain is what you bring to this course and bring your domain knowledge, we will tell you about concepts techniques and tools in the data science domain starting from the mathematics the foundation going up to building the different floors right so including machine learning neural networks, then we will top it off with applications, case studies from the business. So business analytics, how to use these concepts that you have learned in practice on high performance computing missionary, what happens behind the scenes, right. So how does a GPU operate and how can it spawn 1000s upon 1000s of processes together and solve your big data problem, right. So how to do that all of these concepts, tools and knowledge will be provided to you And it will not just be theoretical. So of course, the way to structure it is a theory lectures and then there is practice sessions, right. So when you actually do mini project that is industry relevant mini project that has been put in such a way that the concepts become clear to you, right. So, in many cases you are to abstract out the different things so that a particular concept can become clear to you. So, the goal is that you learn how to use those techniques and tools. And using that knowledge, you will be ready to apply it to any other real world, whatever is your specific product problem, you'll be able to apply. So that's the goal for us, right. So this is the need for data scientists. And that is a gap that we are trying to fill. And hopefully all of you after doing our course becomes champions in your own particular fields, your own particular companies in taking from the proof of concept stage to a minimum viable product, and you will all become successful business leaders and data scientists with that as a goal for us. And there are some FAQs that I think have come right, so shall we go into some frequently asked questions.
Absolutely. Professor, you want to take that up? Or do you want me to set up the questions for you?
So I think I have the questions. Right, so the first question, I think we have answered it, I guess in the last 4045 minutes, you must have understood it by now, right? How a CD is different from other data science programs. So as I said, we focus on fundamentals as well as practice, we have unique features in the program, such as the data stories, industry relevant mini projects. So that's how the course is structured. There's theory on Saturday, and then there is assignments on again on Saturday, and usually on Sunday. So this can change from core to God, but on one of the days of the weekend, we have theory and assignments and the other day of the weekend, we have a mini project that will make you do it by yourself right you will actually do the coding and you will learn that and our faculty as you know are leading researchers and they know the best and the latest technology and tool that is available concepts that are available for a particular task and at the risk of beating our own trumpet right. So, we could say that we bring the best data driven research based teaching pedagogies in the way that we communicate information to you and hopefully that has become clear to you in the last 30 or 45 minutes in our discussion then who are the faculty right? So, as Professor she already mentioned, so, faculty all doctrines right. So, I have AC from MIT, he has a PhD from OBG in Germany, other faculty from Texas a&m University Michigan,
they are in we are put together I have a lot of industry experience working or consulting either with all the major internet companies and even fmcg and aluminum companies, oil and gas companies, I also consult with a lot with startups nonprofits, especially in the environment conservation domains, right. So, we are all variety of industry experience faculty. And then I guess lots of questions that has come in the chat is an ideal candidate for the course right. So, what is the ideal profile of candidates and the only thing that you really need is enthusiasm right. So, the enthusiasm to learn to ask questions and never be satisfied until and unless the concept goes into your head that that enthusiasm if you bring to the table that is the most important thing and the prerequisites are not much right. So, you just need a basic knowledge of mathematics This is as a first year engineering undergraduate level right. So, at the level of the first year engineering undergraduate, that level of mathematics is needed and an interest in programming and the way you think right. So, the computational way of thinking and programming and some knowledge already about what a for loop is what he felt his condition is right. So, how to write all those those concepts must be there. So you must have done a CS 101 kind of course, right. So the first programming class and you know what these basic concepts are, that is required. Now, the higher you are with the knowledge of mathematics and programming, the lesser effort you must put during the course. Right so if you if you start off on a slightly More level of the basic mathematics and programming that much more effort you must put during the module zero. If you do that you will be good, right? So you can definitely cope up with the remaining part of the program. So indeed in module zero, we have experts who will teach you how to use Python. So, if you bring in a knowledge and enthusiasm and you practice, practice, practice, practice, practice, practice practice, I can keep on saying that till the end of the webinar, right. So if you do that, you will be fine. Right. So mathematics and programming are the pre requisites. And we do ask that you have at least one year of industry experience. The reason is that you must know how it is out in the industry, like where will data be used, and it puts you in the right context, right framework of mind to absorb whatever we teach, and immediately put that into practice. And you will ask those questions. So that becomes really, really important. So if you are enthusiastic, and you want to learn, right, you have basic knowledge of mathematics and programming, you're good, right? So that is the ideal profile of candidates and the willingness to work really hard, right. So, hard work really pays off. And if you do that, you will be really fine, we will give you all the concepts, we will give you all the material, you have the lectures that are available, you can re watch the lectures, right? You can play it at point 4x 2x 1.5x whatever is the comfort, right? You have we have live lectures, right? So live lectures for which will be recorded, and then you can watch the same lectures again, and again. That's what I mean. Right. So that that is what it is. Then the other question that comes usually is, what is the value of this certificate, right? So so this is an advanced certificate from the Indian Institute of Science, it's so advanced to say that it is PG plus level, right? So the content that we teach here is the same content that goes into all our doctoral goes doctoral programs, all our master's programs. So it's not just PG level. It's beyond that also. Right. So that's an advanced program that we certification that we offer. And if you want to know I Professor she already mentioned. So the NATO science CDs department annually for our regular MTech and PhD programs, we get over 3000 applications, gate qualified m tech m many qualified candidates, right. So this is the, and we admit, I think about 40. Finally, sign every year, between 30 to 40. Right. So from that ratio, right, so we really want to scale up and bring this knowledge out to working professionals. And that is the value the certification brings. And hence you must expect that kind of rigor from the course. And that much work that would be needed. We will provide you with all the guidance, right? So as they say, you can take a horse to the water pen are over now the horse has to drink, right? So we will take you up to where and if you are willing to put in effort or the I think it's about 10 months course right? So you will really be ready to tackle the next question, which is that how will this program help me get industry ready, right? So we provide you with the knowledge, we provide you with the hands on experience. And through our unique data stories module, you actually will prepare a portfolio to showcase what is the knowledge that you have out to potential employers or to move in the ranks in the same industry? Right. So that's all this is some of the frequently asked questions that we have answered. And now we can take other questions if there are any. Yes.
Thank you. Thank you, Professor Deepak. It's good. You know, you there is on this trial, there is this bring your own project option, where I'm kind of looking at some of the questions saying that, you know, do we get industry projects, etc, you do actually have an option of bringing if your company is willing to share data or you're working on some project. And if you can get that data as a part of the project, or as a part of the learning process itself, you will be able to bring that and do that as a project. So that's something that's that Professor do one request is that you know, data stories are something that's really interesting. And that's one of the biggest USPS of the program. We've been speaking about data storage a little bit, maybe you may want to just cover it up, you know, couple of lines about what that is, because that's something that most you I don't think any of the other programs have in the market as we speak.
So data storage has multiple facets to it. So one is really in communicating what you know, to the outside world, right. So are you able to create these narratives from data or not? Right? So what have you learned from the course and to communicate that in Nice format that enables the audience who are reading. So it could be in the form of a blog article, right? So you may or might read about towards data science blogs, right? So might be in the form of a blog article might be in the form of a video that you produce, it might be in the form of a portfolio that you construct, so that there is no limit, right? So sky's the limit on how you want to convey data stories out to the world, right. So that's what the data science module is all about. Right? So to give you a concrete example, I just find a concrete example. In that time. If you have other questions, we can reveal take up
some of the other questions. So there's a question from Radu, you can send in your questions we will try and answer majority of them. These are the ones which have come in early insights about placements after successful completion of the course is there any time the company has campus interviews, again, there are no campus interviews that happen as a part of the program. Also, this these programs are primarily for working professionals. So the assumption is that you are already working in your company and you don't need placements as such, having said that, you know, as a part of the talentsprint career accelerator that we have career accelerator, you get access to curated career opportunities that come in. And and that's a part of our alumni network, this is independent of anything that ISC does. So you get a part be a part of that. That's an alumni network of all of our programs, you know, be it from ISC made from IIT, Kanpur, I am Calcutta etc. And there are certain programs and there are certain activities that happen in parallel to what is happening in the course module, which is there. So, there aren't any bureau placements as such, but if you really look at it, again, placements is something and I've always been in, you know, the people upskilling space for over a decade now, most people concentrate on placements, My only advice to people is that don't look at the placements part of it, look at the skills that you have, if you have the skills, the job will come to you. If you even if with placements, you know, if you do not have the skills, you will not be able to hold on to the job. So that's there. Is it possible for classroom training for Bangalore candidates? No, there are no classroom trainings that are available, you may have campus visits as a part of the program where the entire cohort will come down, together. But even if you are in Bangalore, you have live online sessions, which happened over the weekend, you're going to be a part of those sessions. And it's just like being in a normal class, you know, you can stop the professor, ask them questions, look at his or her, you know, whiteboard, look at notes, etc. It's all delivered through the talentsprint platform called ipoll.ai. Incidentally, the same platform is being used to deliver the main MBA program of I am Calcutta so that way, if you really look at it, it's a world class program it a couple of IITs are also using the program to deliver their, you know, core PG level programs. So that's their professors. Do you want to talk about one? screen?
Sure. Right. So this is from New York Times, the way they say journalism through data, right? So data stories, which New York Times does. So this is a map, right? So of the total corner cases that are there. So this tells you a story. And the story that it says is that the COVID cases have actually been and it has come started to come down in the globe, right? So all over the world. So saying the story through visuals, and it's very stark, right. So you see this picture, you immediately know what is happening. So how to choose the right tool to present the story or whatever you want to say from the data. So visualization is one aspect of that. So this is another visualization, which is a geo visualization, right? So how and you hover over it, you get some predictions, right? So you get some case, recoveries and all these different different things. So this is one way in which you tell a story with data. Now the other data stories aspect is that if you actually can write a blog entry, right, so if you are the author of this particular blog, then the employer right so whoever is interviewing you, clearly knows what you know, right? So they don't need to interview you at all right? So if you detailed right give with code and examples on what you really know. It gives me a confidence in hiring you, right? So that Okay, I know, you know this much. So if you build a portfolio, so usually artists, right, so they create a portfolio. Either they draw or they sing, or they send samples or they send, if you're an actor is in the side shots and all these things, right. So that is a portfolio they create. So what is the portfolio for data centers, it's the stories that we say, with the types of blog posts that we write, right? So if you can showcase that, and this is where working with the open data that is available, and through many of our mini projects, you will be working with open data that is available that you can share to the outside world, right. So if you bring your free to bring your own project, if you don't get me wrong there, right. But if you do that, you might not be able to share it like this. Right. So that's one consideration that you must really keep in mind. Right? So these are just examples of how you can say the data stories. Right.
Absolutely. Thank you, professor, I get back to a couple of the other questions that we are seeing. And I will I'll request for your help in answering some of these. Swami sollen asks who he hates business analytics and fmcg company, but he feels that he's unable to below he needs to get better at deploying, developing and managing models at scale. And he has transitioned from a transition from mechanical, technical domain management, he will roll seven years back, will this program be able to help?
Yeah, maybe I can pick up the solution. Yeah, so in this program, as we have already discussed, right, it's more about the learning the techniques and learning the fundamentals. And also, we will then also teach the tools that can be deployed in the large scale data on the large scale machines. So it's all about their participants who utilize all these codes and patterns and what they learn on top if the intention is to learn and apply the large scale problems, big data on done in the large scale missions, then the from the beginning itself, they have to concentrate mainly from the module one onwards, how do you how the parallel computer walls, What are penalisation? What is the parallel architectures, right? So if you look at the all the tools from starting with the
airport, or even the TensorFlow, the every, almost all tools are using all the high performance computing computers. So it's all No, it's all about knowing how to deploy these tools on those type of machines, those things will be covered. And then what these participants can do is they can come up with maybe normalized data, what they are planning to deploy, they can come up with a moderate data, and they can do the capstone project on the database. So fancy how they can be deployed on the cloud computing are there in our mission is having already the big computer supercomputers, they can also check how it can be deployed there. It's all about the as Deepak mentioned, right? It's a we provide the basic unfund foundations of the tools, it's all about how the participant utilize and according to their requirement, and then the, the effort, right, so that's how it, so I don't see the reason why they cannot get it trained on the PI for their permission systems with the large scale problems. coming with a clear mindset of doing that, then I don't see any reasons why.
That's okay. So lankesh answer to your question is yes, you can do it provided, you know, you're open to working and you know, we've been talking about that for a while. I'll move on to the next question. Promote has a question at is that what's the difference between data science and data engineering?
Okay, and I think that those parts, definitely we will talk about, that's the very first lecture that will be part of the the chord, right. So there are engineering is mostly about the starting from the collection of data, cleaning the data, and then how to bring this structure on how to get the clean data, from the data that you're going to get it from either from the survey or from the Netflix or the Internet of social media sites, it's all of the data comes with a lot of noise. So it's all about converting the data into your useful data formats, right. So that's when mainly the data engineer comes into the picture. But the data scientist as a whole, it's all about, it's all about framing the problem, right? How about this data? There's a cost and the classification, I would say how how they're assigned pieces are really different from the data engineers. It's mostly about the framing the right questions or right problem out of this data. So I have such a data. The question is what I really need on how can I extract the information from this data? It's not the beauty of the data scientists to Go and collect the data. But on the other hand, if you're in the manager role, you one must know that even the data scientist must know that there are engineering, but that's not really the part of the day to day jobs. So data scientist comes from the take the data, and then frame the right proportion, right problem, and then how to extract this information. And maybe again, the data story is a classical example. It's a part of the data point history. So you have you got this information, your machine learning model, the artificial intelligence game, all this information. But the question is, how we are going to represent how you're going to explain to your boss already explained to the public explain to your clients,
it's all about the data storage. So that's where the data scientist role becomes really crucial to the data engineers. Of course, that is when you say that are all working on the data, but the role changes when you say that one person needs other knowledge definitely needs other knowledge, but the day to day work is completely different.
Absolutely. Thank you, professor for answering that. Let's go on to a couple of other questions. You know, Venkatesh Rama, Chandra, you have asked about, you know, how is the you know, people coming from different experiences and backgrounds? How do we cover that? I think that's already been answered. There is a bridge module that's there that you will start before the program, which kind of helps bring everybody onto the same page Shivani has a question, what's the admission process for the CDs program, admissions for the CDs program are on, you will find that there's a chat link that a new clarity has posted over there, you can go to the talentsprint page, you can make an application. After that, you will have to write a small sob after that. Your profile is then sent to both Professor Deepak and Professor Shashi for shortlisting, they will consider your profile and then it's shortlisted, you get an offer and you then join the program. We are expecting to start the next cohort sometime late April, early May. So applications are on I mean, almost, we have about 50 odd people who have already enrolled by the time we get to the end of this month, we'll probably be close to about 100 people who are there. So there's, you know, I would request if you're interested, you can make an application quickly. So what's the let's look at this. So prerequisites for the program again, you know, many you have this question we've covered that as well. Okay, Satish is this program equivalent to PGD, MBA degree in the same domain, how it is, again, see, if you really look at it. The equivalence of a degree is, is something that you know, that you you probably be looking at the certificate program is not equal to a degree, I'll be very honest, right. But then if you're looking at this as an avenue of upskilling, yourself, this is a great program to do if you're looking at this from a perspective of you being able to do research, etc. And other than that is not there, it is not going to get you into say I need to do you know, additional program or higher studies, etc. That's not the case. And I believe that would be the same for other universities as well. If you are looking at this from an upscaling perspective, this is probably one of the best programs available in the market. Okay, so. So professors, there's this question, I mean, you know, a Shawn has this very long question BA in economics, etc, knows basics of programming, etc, etc. Would you want to clarify this, you know, we have this, you know, one year work experience, we kind of spoke about that. But in for the first cohort, we have had a process where bright young professionals or emerging professionals, so to speak those who have not yet become working professionals, we have given an opportunity to them. So we want to just talk a little bit about that. Is that still on Firstly, and if it's on we want to talk a little bit about the process.
Yes, so, if you are fresh out of college, right, so with just a degree and no work experience, then you can apply and we will have interview process after that. And based on that interview, you may or may not get the offer to join. Right. So that's the process for if you don't have one year of work experience that we say is required. Yes, thank you. So not only have vsee rates of three year graduation rates, if you only have three year graduation then we do require that You have at least five years of work experience in data related or programming it job right so that you have got that experience. Otherwise usually we don't accept a three year BSc course. Yes, thank
you for clarifying that. I was just gonna come to that as well. Okay. 10 non working candidates apply. Yes, non working, I'm assuming you have work experience, but probably if you're taking a career break or something like that you can apply there is, you know, nothing stopping you. As long as you know, you fulfill the entry criteria, or an answer to your question, do we get access to ISC alumni? Not really. But then if you are visiting the campus, I'm sure you know, you will be able to interact with some people who are their professors, maybe you may want to, you know, take a stab.
You will be the talentsprint alumni. Night, you will be fired not specifically iisc alumni but you will be talentsprint I CCDs alumni, right. So that does you will have but other than that, there is no access to alumni network of IAC.
Absolutely. What, okay, let's look at this. Okay, Maddy asked, from a time perspective, how much time on a weekly basis is required to be dedicated to accomplish this course? We get this question a lot. I we should probably put that in the FAQ.
Yeah, sure. So the actual contact hours rate is so so there are two on a weekend ritual course runs through the weekend. So the day one of the weekend will be four on four and a half hours of contact, right. So half an hour of quiz, and then four hours of lecture, right? So four and a half hours roughly. And then there's an optional assignment tutorial session where you can meet with the mentors, work through problems and answer them. So that is another three to four hours. And then there is a mini project, which is again, about three to four hours. So on a weekend, roughly, we are looking at 10 to 12 hours of time that you will spend either with the faculty or with the mentors, right. So that's what we do. And really, the program has been designed in such a way that if you spend that 12 hours actively, right, so you put in quality effort during the hours that you are there, you should be able to get the minimum or the on an average rate, you should be able to perform on an average. Now, now that can depend on you. Right, so from where you come from. And so it's like a running race. And you some people are ahead, and some people are behind, right? So the more you work, the better it is, right? So maybe on top of this 12 hours, I would think like if I were doing this program, I would maybe spend three to four hours more on top of that to like, kind of top it off and close the lid, right so that the information doesn't go out of my head sometimes. Right? So to do that, maybe that's the kind of time that we are looking at. Yes.
The The thing is that, you know, while we get these questions a lot across all of our programs, one thing that you have to do with that, you know, these are just ballpark numbers, it depends totally on your profile, I mean, your learning pace, you may be able to pick things up quickly, etc. So take this numbers with you know, juxtapose that with your learning ability, and the amount of time that you will be able to spend. Samia has this question, how is the learning validated in the program, which is basically probably asking about assessments, etc. Can you just talk a little bit about that?
Yeah, so we have both continuous and terminal assessments, right? So both formative and summative assessments. So we use the best teaching pedagogy of using equal formative and summative assessments. So, the formative assessment that is a continuous process includes activities like online quiz, that you attempt just before the class to test for understanding of what you know, right. So, there will be a quiz on every day of the lecture will be a quiz. Then there are assignments, which are like lab sessions, tutorial sessions, which contains points. So that is also every week. And then you have many projects that you will do in a team of 1010 people currently, right. So that is a team activity where we will actually simulate kind of real life example but of course, abstracted out to something that can be completed in a duration of two to three hours, right. So That mini project is there. So that will also be evaluated then towards. So there are six modules towards the end of every module, there will be a summative assessment, which is like the final exam, which will cover your concepts, which will test for your concepts and your knowledge. Right. So that's the process of evaluation. Right.
Let's look at the other professors we've overshot time, I hope you are available for another five, seven more minutes to answer some of the questions that are coming in. Sure. So
don't get worried by the exams tested. All right. So it's that we offer a certification from IAC. Right? So that value must be there. So to ensure that the certificate is valuable, right? It's not just given out for free like that. Right? So to ensure that these tests and the testing of concepts are there, right, so you see it from that perspective, right. So it's the value that it adds to the whole?
Again, sunshine, this is probably an answer to your question, the last couple of statements that the professor gave, you know, while attempt to interview for data science profiles, you have asked this, what's the value of the certificate? See, it's, again, not about the value of the certificate, it's about the skill that you are able to bring in to the table, I see is just Yeah, it's a validation of the fact that it's a very good program, etc. But you still need to demonstrate the skills. That's what will get you in. So there's this question from
a little bit more, right. So I have sat through placements in IIT at MIT, and I have seen my students go through placement process in IC. And I can tell you and I have sat on the other side of hiring also, right. So I can tell you with 100% confidence that it is only your skill that matters, right? So what you bring to the table is the skill. And that is a skill that we will be providing, I hope right through this course. Right? People who have degrees from MIT and who don't get placed, right? So because they don't have the job industry ready skills that they bring. Right.
So absolutely, we are thank you for that. Professor. That's a good very good perspective. Nursing ma has this question he wants to understand, you know, any technology cloud platform for the hands on sessions? Or will they be expected to work, you know, only on their own laptops, personal computing devices?
No, you're encouraged to use Google collab, we all the lab sessions mini projects will be in Google colab. And for whatever you do in the course, colab is enough. If you have some specific other requirement, you can get the paid option of color. Or you can set it up on your own device. But now colab is really good. So we recommend using color.
Bright, Thank you, professor, so many Shan has this question. He has 19 years of experience, but max fundamentals? Again, you know, he's not sure if he's good enough, or if those skills are good enough to participate in mohnish. You will be I mean, it's something that you will have to take a call on, we've been talking about the prerequisites that are there. He's also based in Singapore, we do have an attendance criteria for the program. Professor, I think it's 75%. Right? Yes. So for all the sessions, you actually have to have that and these are live session. So you've got to, you know, attend the sessions while they're happening, and that the attendance is automatically captured. If you're able to manage that with the time difference in Singapore, it shouldn't be that difficult.
Think about it, right? So I think Singapore might work out fine. It's two and a half. Right? So actually, it might work out better for you then Indian time.
Okay, so we have been talking about it, but pushpendra wants us to elaborate a little bit more on the mathematical concepts in the program, what levels would be covered? What would be covered an optimization?
Yeah, so in linear algebra, we will be dealing with vector spaces, matrices, tensor products, Hadamard product. And in optimization, we'll be dealing with first order methods second order methods, Stochastic gradient descent, Batch gradient descent, Adam, RMS prop saga, all these optimization techniques. And in probability, we'll be dealing with the mutual information, entropy, all these ideas, and you must have done a course in probability linear algebra and Porter optimisation but optimisation is not needed at first year, undergraduate level, right. So at that level, you must have done a course, if you have forgotten, that's okay. We will refresh it in the module zero. But you must be comfortable with it. Right. So you must have heard about it at some time in your own career.
Right. Great, thank you professors. Last probably two questions. And then we'll be done with this. By the way, I've opened a poll if you you know, if you can vote on that as well, that would be good. That would be good feedback for us. Prince has this question, do you have data science industry professionals collaborating during the course, it will definitely help. In addition to the panel of faculty members, I take my personal example, I am currently undergoing an AI program here and marketing programs from I am Calcutta. As a part of the program itself, we have about 80 odd people. And we come from around 8090, not industries, etc. We have on our own, we have our friends, all of us are very senior professionals who are doing that program, we have arranged for at least 18 to 20 of our industry leading friends to come in and address our class outside of the class hours. So that is something that is always there. It's not there as a part of the program. But we are interested. So collaboration is to a certain extent dependent on the students itself. And there's always a possibility you can use the platform, the platform talentsprint platform is used extensively for by us for all of the external stuff. So if you're really looking at it, there's enough collaboration that is variable anyway, have, you know, data scientists, many of the people who have joined the program are working as data scientists already, you will be able to collaborate. So it's it's not a question of whether the program itself has it. In this day and age, even if a program doesn't have it at at any level, it's always possible to you know, get that in place. And thanks to technology, you can do that. So, where Okay, can you push pendu if you have questions, which have still not been covered, you can reach out to our liquidity liquidity, we request you to share your email address and phone number. You know, you can reach out to her she manages the admissions for the program, she'll be able to answer questions. Okay, so I think
so, there's a question on from learning content perspective, how would you compare the Certificate course with online data science, Master's Course from, say, University of UT Austin. So the content is at the same level as any other online data science Master's Course. And it is definitely I can say for sure of the level of the course. And equivalent courses that are is offered at MIT. So that's what the content is about. And that's, that's it.
Yeah, I'm Thank you so much. I think it's been a wonderful session. We've overshot by almost half an hour. And thank you for being such good sports. I know both of you are very busy, for giving us the time today to answer your questions of people and thank you to everybody you know who stayed back for the session. I know we overshot time, you can reach out to us at ICA talentsprint.com or under pity has provided her number reach out to us. And if you are interested, you will have all of the details available as an email which will come at the end of the webinar, which will come sometime between tomorrow and day after. And we look forward to getting to know more people into the class and seeing some of you in the class that is going to start very soon. Thank you, Professor Shashi. Thank you Professor Deepak for your time today, and I look forward to having more such sessions.
Thank you and all the businesses around the participants. Hope to see you all. Thank you again. Good night, professors.
Watch the entire interview here https://youtube.videoken.com/embed/pOTxi55nZzM