Get Ready for the New Normal with AI/ML Expertise
Okay, good morning, everyone. Thanks for joining us on a Saturday morning. I think as a little said, We are at level and start already. We have 50 odd people on the session we expected fukrey to 60 people. Thanks for joining us on a Saturday morning. It's good to have you hear. My name is Sweden's Chief Business Officer at talentsprint. I mean, I've been talentsprint from almost a date started 10 years back. I have along with me, a shokan, whom you can see on screen what I will do is stop the screen ashokan
I'm here. short intro will be useful.
I've been in
it programming and software development for a lot of time, nearly 40 years now.
Now. I also have
been one of the early users of both Linux and fighter. I built a Python system where Back in 96, and I've been a Linux user from about the same time
prior to charge of
designing and developing programs, and I also beach part of the AML program, I also beach fight they're not
part of the course of a DNS on
that they have backup yesterday though.
Thank you. Thank you. shokan. So great to have been ashokan had some other shuttle but I pulled him in this morning asked him to join me today. Thanks to his experience and expertise, I thought it will be great to have some interaction with all of you during the day and hence, requested him to join me today. Okay, so, in terms of format of in, let me just start. I'll just spend a few minutes just a couple of minutes to three minutes to introduce talentsprint. So there is a context to what we are doing. Today, follow it up with what we are seeing as new normal, I'm sure the lockdowns and the COVID and the way things are a lot of things have changed in the market. And I'll probably share what we are seeing and make it interactive. We will require some of you to share your views, and then follow it up with what should a tech professional at this time to to go to the next level and what are the things that makes things be ready for the new normal? Okay, so with that background, let me just move into the short tech.
Are you able to see my screen? Maybe a chat message would be useful? confirmation would be useful. before I proceed, we can see that. Okay, are you able to see good? Okay, so as I said the agenda, I will spend a lot not more than three, four minutes on about talentsprint But we will spend most of the time on the second one which is the primary theme of the event which is a new novel. And the third is AML expertise and what do we do for tech professionals in this area? About talentsprint just as a background talentsprint is a 10 year old organization. And as a company, we are exclusively focused on deep tech. We can call ourselves as a tech player. And we are focusing on our deep tech programs both for working professionals and graduate students. And the programs are in areas like AI machine learning data science, FinTech, blockchain, cybersecurity, digital health, robotic process automation, digital process, automation, etc. The next one is about all this problems on top partnerships. We will talk about it in the next slide. I don't want to bring it up here. And this company has been launched and the on managed by a team of professionals from the industry as you would see.
As your chewed up, shunted off ex CEO of society is IIT Madras graduate areas of Michigan top rate started his career with IBM Watson labs, myself fire more from the finance side but of heavy user of technology. During the finance days I used to be the country head for a group called Carvey group before getting into the tech space ashokan introduce himself. Geeta, who's our CTO is IIT Kanpur grad debjani again now is a professional who worked with Microsoft for more than 15 years and now drives our partnerships. Gurpreet x Honeywell and she runs the HR and, and as a company, we have been lucky enough to get marquee investors to invest in the company. So, just another one minute on the slide and one more slide. Before we get into this thing. digital health and imaging is a program that we launched during the lockdown and this is what the India's number one technical institution Indian is Have sex. All of our programs, we try to bring in top partnership with top academic institutions or global corporations in order to make it interesting both for working professionals and graduate students. Most of our programs for working professionals are for the executives, while most of the programs are from corporations or for graduates. So our current program recently launched the most recent program as digital health and imaging. And we have been running a program last three years with AAA to Hyderabad which is India's number one machine learning lab and research organization on AI and machine learning and subsequently, we launched a program on blockchain. At contour, which is the third brand that you can see, we have done the program on cyber security and cyber defense. I am Calcutta one of the top management institutions. We do programs on FinTech and AI powered marketing and IIT Hyderabad on yeah and emerging technology. with global corporation, this is largely for working among college students. With Google Mountain View us we are doing now probably the largest women diversity program called Women Techmakers engineering fellow. In fact, yesterday, hundred people graduated out of the program and 150 people for the next year enrolled for that program yesterday. Pay guys top digital Process Automation company. Again US based and we are the exclusive implementation partner for the program in India in colleges. Same thing we are doing with automation anywhere and blue prism. Interestingly, as you would see, all these programs are deep tech in nature and as a company, we really focus on that and we keep adding multiple programs in the space. In the meanwhile, I see some hands raised. Let me just get there and because your Please go ahead. If you want to ask something You can unmute yourself, because
I am because there is a lot of background noise. Do you have any question? Okay, let's wait for her to get going. And so interestingly, while as a company we have always been focusing on deep technologies and I think that three, four years last three, four years the focus has largely been on that the current situation of COVID, the lockdown the global economic impact of it, etc, etc. I think the world has changed phenomenally in the last three years. I'm sure each one of you are experiencing it. Right. Maybe let's start with something very, very close to our heart. I mean, I am working from my home studio here at Hyderabad ashokan is In the studio working from home as a new normal, right and every company I mean TCS was quoted extensively for talking about 75% of the employees will work from home by 2025. Yesterday, IBM came out with a news that they are cutting their office space by 50%, probably indicating that outdoor workforce will work largely work from home. Google and all other tech majors across the globe have worked on it and I'm sure each one of you are experiencing it. Infosys added
the same yesterday's to them.
Oh, is it okay. Yeah, they
said 50% of the people. They expect only 25% of the people to be in the office.
Yeah, and and as talentsprint we have been working in office for the last three to four weeks when it was allowed to work. And I think we are on an average working on 10 to 15%. capacity and everybody else is working from home and they are able to deliver. For example, when this happened, all our programs completely went online for delivery. Each faculty has a laptop and a tablet and, and is delivering the classes from their home internet at this point of time. This happened, the work from home happened on one side, and that's largely for the HR side of the organization's. But on the other side, the extreme automation as I would call has been no rage over the last four years. Right? all companies are going towards automation. And based on the recent McKinsey report, which said in 2018, there were 39% of the jobs are automated and it is expected to cross 50% by 2022 or 2023. And there is another recent report by Bain which says that 5050 will cross very, very soon. Right, and I think it's very clear. institutions and organizations are completely looking for going digital artists. Part of
time. So, with that being the case, I mean, we thought we should just put up what is getting automated? How does it affect most of our work that we are doing today? And what should we do to the next level? Right? With that, let me just go to the other slide, see automating processes automating production have happened 10 1520 years back, what is currently automating getting automated or very different, and that's thought I thought we should bring those in. decision making is getting automated. Right? I remember I am largely from the financial services field, my first job was a credit officer in our bank, Citibank, to be precise, and as a credit officer, I used to get hundreds of documents, review the documents and get the process going and probably take three days and of course, my sales team will say see that takes one week, and we never can say whether it is three days or seven days yet but that's Used to be the turnaround time and today it happens in seven minutes or three minutes on online platform. Customer service is completely getting automated today chat bots are everywhere and stuff like that. And technology that's driving all this AI machine learning deep learning, right? Interestingly, blockchain is come out as a solution for automating trust. Right? automating the cybersecurity is through various cyber defense processes that are in place process automation, while most of the physical processes have been automated, now, many other digital processor also getting completely automated thanks to the RPA and the TPA, that are happening. One is process level which is RPA. And one is enterprise level, which is digital process automation. And we also see new industries I mean healthcare have put I'm sure we can put education DeRose right. Education is also getting automated. So what is currently looking at and that key takeaway I am taking out of whatever we are seeing over the last two years in general and three months in specific is that I think the future of workforce is going to be very interesting. In the past it used to be skilled and unskilled. Then after that it was knowledge prevent white collar and others. Now, I think it's all about people who will work on creating algorithms and those who will work for such algorithms. ashokan would you like to add something out here?
It's a it's been a dynamic which is moving quite around for example, I started coding sometime in the very early 80s. And from then on, we have been threatened with, you know, programmers only required there will be programs to write programs. That hasn't happened but what is instead happened is an enormous number of tools have arrive in areas around software development, software testing, like that in every field algorithms are adding to the productivity of the people replacing certain tasks and in the process bringing out bigger tasks, the solver is continuous process where what is happening is what was considered a very tough problem yesterday is now a solved problem and then it is no longer interesting. So an algorithm can do.
So, there is a chat saying that your voice is not audible. Can you just take it off your mic, please?
I could hear you. Yeah, please go ahead.
Yeah, see, from the beginning, there has always been talks about something a lot of made everything out of everything. But what often happens is processes which are Difficult to do become handleable once automation arise and and stuffs which we never thought about doing nowadays are okay, we'll attack those. So my, one of my colleagues always used to joke, the problems which were considered difficult and the enterprise level have become exercises for the students in 10 years time and program. So in that sense, yes, we are continuously working on algorithms. And once those algorithms become mainstream, they end up people start using them, either they work for them. Yeah, that's a little scary prospect, but definitely, they work with them. And we are working our algorithms if you are looking at someone collecting information based on what KYC is needed, you are working for someone. Yeah, in that sense. Yes.
Interesting. Good, thanks. ashokan.
Okay, so when we really looked at it and then say that I mean, in the meanwhile, I'll probably add a few data points that we did recently, right. During the month of March when the lockdown started during April and May. We reached out to our partner companies who recruit talentsprint certified talent, some of our programs, right. And one of the things we saw there was very interesting change. I mean, every year we take a take up update on what they are really looking for in the upcoming here, and what kind of skills they look at, because when we do our program planning for a year, we take that into consideration. And it happens generally every financial year. Right. I mean, this may be a little what you call irrelevant to the experienced professionals. This was specifically done for graduate people, but the trend was very clear visible. 2018 2019 2017 we had seven To 80% of the people requirements were on generic skills. Right? I have some people facing hands and chatting saying that the voice is not clear. Is it still so?
Ambika Can you just unmute yourself Intel or mohnish
Hi, Manish Sharma Can you unmute yourself? I would like to take your inputs before we proceed.
Manish you can unmute yourselves.
Nish. No. Yeah. You're on unmute, please go ahead.
I think his microphone needs to be set up.
Man, if you need to check your microphone settings.
Yeah, we are able to clear I think there are others chatting saying that the line is clear. Let's let me be having I also switched off. Oh, okay. So let's just check
your settings. Okay. Okay.
So I will proceed. So one of the things when we did this survey was very clear. I mean 2018 19 2017 1819 the companies used to tell us thing that was get me people who can basically solve problems and can understand and to coding in one particular language and popular language. It's used to be Java anyway. So if they can code in Java, and if they can solve problem that's through aptitude and communication skills, be in good, we will recruit them. And 80% of the requirement of that and 20% used to be generic serving specific skills, right. Do you have anybody who knows by And specifically or do you know to have anyone who has done something on data sites that used to be 20% and when we did the current survey, which is 2020 the tide has changed 80% of Okay. Let me start with basically the number of job requirements came down by 30 to 40%. Earlier What if they said 100 people are required now, they were saying 60 to 70 people are required and out of the 60 to 70 people 80% requirements are not specific skills. Right. And we call it deep tech skills, right? And this deep tech skills are brought out here for you to really look at, I think, majority used to be on here machine learning data science, deep learning, and in different words, right NLP and computer vision kind of combinations used to be there. blockchain was there but limited so that used Be the demand. The second kind of demand used to be digital process and robotic Process Automation As you know, there are at least four companies which has come out in the last three years as a global leaders in the robotic Process Automation right you have automation anywhere blue prism UiPath and digital Process Automation pega has been a leader for long. So, all these companies skills required for these these practices that increasing the other area are the skills I will leave it to ashokan to explain. Yeah.
See Yoda here the communities which are interested in some form of tracking professionals, like the OB run surveys on languages, what they are using, what they are learning. They also run things like what they like to work on how Like, how much they like the language they are working on usually Java's around the top. And item, the last two, three years, this has significantly changed. And now we find, for example, this year there are the very new entrant in the language which people love to work on and people are wanting to work on is changing because of two three reasons. One, people find very interesting things that are happening in those languages. Two people are also finding that they are solving existing standard problems and a much better way than before. I'll give you two examples. There is a language called rust, which is not really getting a lot of traction, and which is in our code, which is in a typical system utility. It's a new C and we also have a language which was designed and innovated by an Microsoft are TypeScript, which got a real boost in its usage because Google started using it in the angular framework, one of the very, very few rare occasions in which Google and Microsoft have agreed on something. So you find today a very large number of people interested in web development using TypeScript. So that is the machine
were the languages and associated tools.
I managed to you have a question? I mean, you raise dance. I have unmuted you. Manish
Yeah, sorry, Ashoka disco hurt.
Yeah, monitors are the areas which are
the when we say niche areas or new languages, these are things which are getting a lot of traction, a lot is relative. Obviously, it is not as high as
there's a significant amount of movement in these areas. Hello,
yes, Manish Please go ahead.
I read in a smile glazy buddy can I do in
Google Yahoo, Google and other info companies
know this session is on zoom so sorry I don't think I'll be able to help out on this.
So Sam, can you just take care of this requirement?
Shall shader okay.
semana Gopal, you can go ahead you have raised your hands you have been unmuted.
I mean, you'll be the TypeScript. I just want to explore
Yeah, well, the voice is breaking in the in the zoom call that I want to do.
Okay, it's breaking is it? Yeah, pick I'll take care. Thanks. Okay. That is someone else was ciphers as for base design. I'm requesting him to speak
I think it's mortal. Some challenge with sound. I don't know, many people are saying it's clear and you're finding it difficult. Yes.
Yes, sir. Please go ahead.
Yeah, so I'm working in aerospace industry and I'm looking for AI na ml application in that so like planning to implement a project in the quality assurance area. So can you put some light on that? Especially?
Sure. The next 40 minutes is about specific AML related question. We just wanted to give a backdrop. We will take it up probably one more slide and after that, we go into discussing specific cases. Saif Is that okay?
Okay, sir. Thanks very much. Okay.
Yeah, good. I think let's, let's go to the next one. And the fourth area is cyber security and now that everybody has started working from home, I think the risk is no Handling extensively and organizations are investing heavily in ensuring that the security is maintained. I'm sure we are hearing a lot of things about security happening over the last few months, right, be it on data privacy, we are on security, we are hacking and stuff like that. I think these are the areas we really believe based on our research on companies and the survey with companies as well as what we could see in the secondary sources of what we are hearing in the newspapers, what the markets are doing, etc, etc. Clearly, the institutions or the organizations are going towards extreme automation. And extreme automation is currently being backed by deep tech, which we saw just now in the previous slide. I would take for either decision making customer service marketing, I mean, I'm sure five years back, we would not even thought of it but we just recently launched a program on the AI powered marketing right It's a area that marketing is being changed extensively right healthcare is being changed extensively using AI and machine learning and IoT and related areas right so we have seen a huge surge in the areas where Intelligent Automation is happening and and i'm sure with the way the economy is way the current situations or the overall demand and requirement of human resources are not going to increase as much as it used to and maybe decrease depending on the organization's and their business performance. Right. Okay, I show Can I think there is some challenge in your videos so he probably you can switch off the video and
as he shot the video,
yeah. So while this is happening, I mean medical side we have talked a lot about I mean, very interesting, the fastest growing area right in the world. Today's telemedicine right. The second area I mean, of course that are topless video conferencing systems and video collaboration system followed by telemedicine and the One other area, which is doing extremely well as a tech specifically the K 12. Education. I think globally, all international universities have completely shifted online for the next term right. At the current term, they are not definitely going to deliver in the physical class bringing hundreds and thousands of people into the campus, I think across the globe is no no at this point in time. Harvard has done away with SAT scores to entry all big institutions or finding new ways of doing things more efficiently and effectively. Closer home in India, I am CSB. IITs have already completed the last cohort. We have been working closely with them and because many of them are exploring our platforms for their delivery, right now. We understand that definitely going forward, the education is not going to be what it used to be till four months back. Right as talentsprint I think ashokan can throw light since we moved a lot of our programs online and how things happen. ashokan would you like to add? Yeah.
I in the beginning, almost quite a few of the programs are not planned to be online, but since the lockdown happened and there was no way people could be asked to assemble in a single place. We had to take paths and then quickly rejigger systems and move on. And quite a lot of people found that very, very interesting. Some of them of course, felt and did say that they joined only for the classroom experience. They would like to take a deferment or transfer which of course we accepted, but a vast majority found that It was actually a lot more convenient, even though they had not thought of it that way. And this was something that they kept talking to us as we tweaked and changed the number of sessions, the duration of the sessions, the timing of the sessions, the nature of coverage, all of which underwent changes in all the programs we were doing. But it came through clearly that barring a very few specific people who had particular reasons to want the face to face experience in a classroom, a majority of people who hadn't thought about it either way, found that the online sessions were extremely effective and were giving them things they didn't think about. One person said it is a lot easier for me now to work at other times because I mean, we're not traveling. The traveling time has saved me a lot of time for me to prepare for the classes. Things like that, which we hadn't anticipated, there are quite a few benefits. Of course, we did find that the amount of continuous time they can work online is less so we had to change. From a pedagogic perspective, a lot of things changed. But we are more than pleasantly surprised to find the majority of the participants were very happy and found it engaging. And we we often were joking about, this is so good. The only thing that is missing is the lack of ability to go out at the end of the day and have a coffee or a beer together. That's the only thing we are missing. Otherwise, this whole thing is wonderful.
Yeah, I think also the fact that as talentsprint we have been doing many other programs, which are 90% digital and 10% physical came in handy for programs like a machine learning which we do with reply to Hydra, which has historically physical in nature, going online. Am I right? ashokan
Yep, that is our platform. has been designed for online delivery and we had specific courses which are designed for that. So it became fairly simple to once he understood the pedagogic adjustments required to move them over to the predominantly online platform for courses which were clearly intended for the other audiences. Quite a few people raised their hands. Maybe I'll just ask him Ah, yeah. Okay.
bottom bar alright. People that are sorry.
And too far away.
You need to offer
discount. You can unmute yourself, please go ahead with your question.
mallikarjun you have been unmuted you can speak one of you can start Please go ahead. Have you ever been allowed to unmute, and speak?
yofoto. Please go ahead.
Yeah. Good morning all. Actually, sir. I'm doing my research in fuzzy queueing theory so that I can relate some algorithm with the AI.
Okay. ashokan It's for you that will be very specific. I'm not definitely in that level of a mobile search. Can we take it offline? I'll give you know, people who can talk to chat at the end of the talk and
then go Paulus. unmuted welcome. come on
you're unmuted please go ahead with your question.
You're unmuted. Do you have a question?
I think the first one of the practical problems and
Yes, go ahead.
Yes, no, she can't please go ahead.
Hello, yes, go ahead please.
These parameters are held by the cooperator Hello sir. Yes, yes, this program is
the lectures are delivered right reply to professor's and the set of is going to apply to you. And of course talentsprint delivers all the practical aspects. Okay, so, okay. Okay, so now I think there are specific questions about the program ashokan. So I think it's better we go into that. Yeah.
so maybe I will start with a few questions. There are some on the chat and some of them are coming in. I will ask you questions. And from that, we will go ahead. So I checked in more than 2300 people have gone through this program, and I am sure individually you have interacted with most of them. Yep. Right. So in the last two and a half, three years, so that what has been your experience? How do the working professionals learn differently than graduate student because you also run those programs? Yeah. For graduates. So
when we talk about
when we talk about
working professionals, starter homogenous Yeah,
can Request me she counts to mute his mic, please.
Yeah, all right. So the gra the working professionals is not a single homogeneous group, we have
sections of them depending on the number of years of experience they have. Another slicing would be based on the specific areas they work in, and then a third slice would be based on what they are looking for out of this program, too. We have found that all of them are first and foremost, the biggest changes are the adjustments have to come in the fact that many of them had not sat through our learning session of this degree of intensity for many years. So after the first few cohorts, the process in trouble it and we sat together and designed a module zero precisely for that reason. And we use it to let people slowly and somewhat gently lead it into the rhythm and cabins of regular classes, at an ends coming prepared for classes, doing the exercises and so on. Okay, that's the one important thing we found, which we addressed by way of the content and the design by adding a model for it. The other thing is, like I said, depending on where they come from, they have different takeaways. And people very quickly identify what they want to do, and work on that. I'll take a few specific examples so that what I mean by that is clear. We have had, for example, a person who was a director of QA, in multinational so I asked him why he was doing the program. He said, Look, everyone is talking about machine learning and data science. No one from our side knows How to do QA on back. So, it is going to be something which we have to figure out. And as the head of such a group, it is for me to understand so that when my, I sit with my people in my group, we can together craft solutions focus. So that was his expectation. So, he did contribute a lot he, in other year, interesting areas of the projects and so on. But this was his reasoning and he did we are in touch with such people, we talk to them, are they implementing any of the ideas and please come and give a talk. They come and give a talk. So, he's very happy and he said, I can't share what I am doing because of organization requirement, but I am finding it very useful. Another person was a pre sales person from service industry. So, I was really surprised I asked him, Ahmed, why are you doing this? What's the purpose he had a very brilliant and To say something he said I want to have intelligent conversations with my developers and with my customers.
Green ended up writing a lot of code and he was surprised at it himself more code than he thought he will do. But yeah, that was a side effect, but he really was very clear water. And the other end of the spectrum, we had a person who was running the maintenance group in your physical prospecting type of company, so he could use the techniques for handling the scheduling of preventive maintenance, etc. And he often comes to talk about with experience, you Joe clearly said I had spare time on my hand so I came attended this program. I could see where our water the low hanging fruits I could exploit. Now I have a 10 interns working under me doing different things. So our experiences the working professional, There's been that very many of them come with an idea of I don't know exactly what this is and I think it can be used let me figure it out and they do end up figuring out a lot of these things.
Okay, good. So, that brings me to the next question there are three different types of people who have entered the requirements are different etc etc coming into the program. So, one of the things that I have always been amazed about is the way the program is structured ashokan right. Generally, I look at any program that talks with foundations on mathematics foundations of programming, then going into how the algorithms work and implementation etc. But I see the first module of this program talks about a problem definition Ah, see, so, why is it so and why is it important?
books are written for
master's programs. Books are not written for practitioners, okay. Let people do more. Basically, a very vast majority of the textbooks are written for master's programs and for people working on the field. We are almost always as working professionals looking at how can I use it and that's a very different perspective. And the second most important point, probably a unique point about cancer a unique and important distinguishing point about a IML data sciences. It is actually domain agnostic. It is like the very early days of computers, people used to ask Can computers be used in accounting can be used here, can it be used there? The answer was of course, you had to figure out what is the problem to be solved was it cost effective to be solved using the computer and or how to formulate the problem as a computer problem. Your business problem could be that I need my p&l statement in one week, or I need to reduce the time between the invoices. The money material is applied to the invoices raised the moment I formulate the problem and extract the equivalent systems problem out of it computers could be used their ml is very, very similar. And this is where most of the people are going to be working. Because believe me, no one is going to be writing a new machine learning algorithm. Okay, let me know in the field who's working currently or who's in the industry currently is looking at that if they are they will be joining a research organization. So, we need to be able to use the algorithms for which the first step is figuring out how do I formulate the problem in my industry, my domain, my company, so that it is amenable to the algorithms already in established use and understood by the machine learning artificial intelligence community. That is the idea. Okay,
that's good. So I'm finding out what the You will go into formulate
the problem How do I formulate my problem for example, I could say I am worried about like I would put this go back to the example of the earlier persons he said we have a large number of mere maintenance, preventive maintenance scheduling, and that exists in 30 different spreadsheets since some eight different departments and they are always not agreeing with each other. So, what he wants he decided the problem was gathering information available in one interesting place all over the place. The physical equipment has a boiler plate which is affixed on which these things are punched or written by paint. So, he simply put around for interns to go photograph every single piece of equipment particularly the area in the machine where the last service date the next service is due date etc ever written. The machine serial number was written he asked them to walk around the digital cameras take multiple photographs. Back, then you used image processing text, OCR type of algorithms and then connected it up and all he had a single spreadsheet, which was absolutely correct, and became the master sheet. So, here the problem is figuring out order the problem we solve, and which is where the domain expertise of all the people in a previous slide you had said, domain expertise is very useful this where it is useful, what's an interesting problem to solve? And how do I solve it? How do I formulate that's the way
it works. So
since a structured
or working professionals for practitioners we call it it's structured later on a success. Look through the some of the chat questions on
chat. And just taking the chat question now. I think Satya Priya Mishra vasagle are algorithms important in datasets?
Yeah, they are no one could say ca ml data science actual definition Are coalescing a lot in the last two years. But yes, one could loosely defined data science as a play day algorithms which are presented as some uniform for human decision making. So data science could be thought of as applied algorithms, but presented in a form for human decision making. So, yes, good
applicability of AML for someone who is in information security, this is
a very interesting election Nexus. That's a very interesting Nexus. It would be closer to a research area today. But there are some problems in that area, which are well understood. For example, if you're looking at all our ability analysis, today, there are tools which are trying to do a static analysis of code with ml algorithms to figure it out. Similarly, there are tools which are built with ace algorithms to analyze log files of web servers, for example, to understand the nature of attacks and so on. So some of the sub areas are well entrusted, well understood and worked on but others are in the research territory where information security is concerned. I mean, the nexus of AML and information security, both areas are reasonably well understood. There is no question of shake Mama raffys question again, like I just said, It's for you to decide cancer diagnosis is almost it's a question of what is your raw data you are starting with? What is your raw data is and how are you modeling which type of cancer it will depend on this but the underlying algorithms are similar whether you are classifying apples and oranges and whether you are classifying cancer or no cancer. The underlying algorithms are nearly identical. Yeah. No, we are not covering specific company related product, no SAP lonardo is not covered.
So can you just at this time say what is covered ashokan so that it's easier see, we
are focusing on open source tools. We use pytorch we use the cover areas of deep learning and classical algorithms, deep learning neural networks. We cover a little bit of NLP and bit of computer vision. Okay, you don't cover any specific for example, ma MATLAB The company has its own tool chain. IBM Watson is a separate tool chain but we don't cover any of those. We cover the open source more widely used tool chains.
Okay. That's a question on margin engineering from geo right are working for four years. Being some studies in machine learning for the last one year in a vacation is there a demand for people like Who do about 90% mechanical engineering like poke at the age,
I talked about someone who precisely is in your situation who was doing preventive maintenance of equipment, like he is ending up today leading data saying machine learning group in his company because he came up with this. He your biggest plus as a marine engineer would be understanding the problem that needs to be solved. And once you understand machine learning a data science to be able to formulate the issues that face you in this language, then yes, it would be useful.
Here's a question on career transition. So, this is probably Yeah, this is a program where you get expert knowledge from the program and historically we have seen that the 2300 people I mean, I think we are getting into the 16th cohort now 15 cohort now, 14 cohorts have gone through, people have found it extremely useful for growing their career inside their organization or starting something on their own or they have moved to New organizations or at different companies, they have worked for the end of the day triple it just as a background, I will tell you, AAA, it is considered. I mean, it's the number one computer science research institution in the country, it has the largest number of papers, quality papers submitted in global research areas and both computer vision machine learning and NLP, they have a leadership position in India and across the globe. So, this you will be learning from the research experts from there and the practical inputs coming from talentsprint and our practitioners. So, as a program when somebody does this program, historically, we have seen that they have been able to achieve their objectives, the objective with which people join the program or different as ashokan said, People want to learn how intelligent conversation with the customers for instance, ample, or they have a real problem, because I would put it this way, if you have a real problem in the organization which needs to be solved, which will make you extremely valuable to the organization, right, or you want to do a start up on your own, this program will be extremely useful. Because the approaches identify the problem which can be solved, define it in the machine language,
the language of machine learning the language of
vision, machine learning, and then from there, go ahead. And we have seen a lot of interesting stuff. Some of the students have worked with us and created some solutions for us. Like, for example, one of the questions I say is I get thousand people coming into the site who is going to be a real prospect for me, right. That's the question from marketing person, right. And it's being answered using machine learning and people have a bit i'm sure companies are willing to invest extensively for a person who can do that. Right. Okay, so great now, career transition. Can we use machine learning in stock markets?
You can use it everywhere. He says like as fast as you can use computers in yecs. If the answer is yes, you can use MLA next how to use R. That's a interesting question. And that is likely to be either already done or you will be the first person to do it Prosser. Java loves to put it that way. If you are going to ask how to do X, either it means you haven't. You don't have the information about people who are already doing x, or you are the first person doing x.
It's a billion dollar idea.
So backing domain, yes, he said
banking, financial services insurance is probably the top users in addition to e commerce, right. Where the numbers are, I mean, they you say this extensive right is about is it mandatory to have experience see we prefer people with one year experience but highly interested people who do not have experience and have other objectives like going abroad for further studies etc. Want to get into the program, we are able to
no longer have people exception or three people are admitted like that.
How do you say MLM b2b market space? I mean, does it can it?
That is exactly what you should be able to solve for yourselves because it's again, like asking how to use computers in b2b marketplaces, the purpose the level of abstraction is similar it can be done. The issue is someone who has worked in the b2b marketplaces has to look at a model and say, Oh, this is the toolkit. Now, let me see how I can use it for my job.
This question is interesting blockchain as the trusted approach for this now, one question before that,
yeah, statistics. No sir. One thing Kumar is this program is not for developing a model. algorithms. It is a practitioners program, you use algorithms, the existing algorithms to build the solutions. Yes, a small number of people in the program use it as a stepping stone to do research. That's a small number. But yes, this is a stepping stone further. So if you want to become a practitioner, you don't need expert knowledge, but you definitely need understanding of statistics and understanding of mathematical abstractions. We don't talk about again, Azure ML the covers along with the specific toolkits of Amazon and Azure known session.
So, there is a question about from yadi about AI based marketing. So cdca machine learning program that we are talking about from purple it hydropower is a program for technical people who can use their algorithms in solving problems in any of the domains including money All right, as talentsprint we have also launched a program on AI powered marketing with I am Calcutta. That is a program for marketing professionals on how to use AI and machine learning and other technologies. That will be less technical, more management and decision making. This will be how you convert a problem statement into solution through using algorithms. Right, because that answers your question. There are a lot of queries that have come in the middle. I think there are some people who have asked for Praveen, I have unmuted you can you just go ahead ask your questions.
I mean, every
okay. We will continue to go ahead on the chat because in the meanwhile computational fluid dynamics, how can I apply machine learning.
Dr. ramchandran, who's a professor of aerospace in Bombay, I don't
know that it is applied, I would check on that.
Again, I can talk to him and give you a answer.
That's the start from scratch.
We assume you know programming. You have done some programming, we assume you have done some math and statistics. We conduct bridge courses, both in programming and then to a larger extent in programming and to a smaller extent in math for people who study but feel rusty, so there are some bridge courses in that sense it starts from scratch. But I have never done programming. I didn't do math anytime in my studies, probably you won't be able to manage the course.
Yeah, there's a question from cyber been about ml and life sciences are shocking.
Again, it's in Milan Life Sciences for a lot of work is happening in C. Let's, at a fundamental level, what is machine learning doing looking at data, looking for patterns and trying to identify potential patterns which can predict what will happen later. So every field This is applicable stock markets is the same thing. Preventive maintenance the same thing, when is a software failing or when is your server failing? Same thing you have the logs of the last five years when the server crashed. If we're able to detect a pattern in what was happening before it crashed, then it is amenable to a ml attack. So in that sense, in any area where there is existing data gathered, and we believe that data contains the information necessary for us to draw inferences You can use a and m.
Okay, there is a question about what is the difference between mathematical modeling and statistical modeling?
I wouldn't really know because
I'm assuming she can't I don't know your question you keep saying please answer me please answer me. I don't know your question. Can you type your question a second time rather than saying please answer me four times.
Okay, that is Kumar who has raised his hand I will request him to unmute
Jai Kumar, you can unmute yourself and speak.
Hello, hi Kumar. Please go ahead.
Physically, I'm looking forward to Olympus technologies. The main question is like when I look at the program, it is offering Three months, actually, around three to four months or so, yeah, four months. But when I look at in other programs in other Institute's they're offering around one to one year to 18 months like that, but is it possible to become master in this technology in span of governments have fundamental doubt actually, because I'm very new to this program these technologies. You mean, somebody should be already Elvis needs technology so that they can expand their knowledge or
you need Let me repeat the prerequisites our experience in programming is rarely enough to switch to hands on coding for some time. good understanding of mathematical obstructions and experience in the industry to look at Industry related problems, these are the three. Now, at the end of three months we have had people who have started startups who have gone on to head data science or machine learning groups within their companies okay, because they would if certain you are a person with 10 years experience, you are going to sit down and write code for it, but you want to know some understanding of how it is done. So, at the end of the program, if your organization is looking at similar problems or can profitably look at similar problems, as a senior officer of the company, you present a plan to say these problems can be addressed here is how we hire some programmers from this area here is how we some hire some data scientists of this caliber and put together a team to solve the problem. So this is what a senior person would preferably take out of the program. If you are a person with one or two years experience and you still want to write code and things like that do go in and say, you know what I would like to move to our data science group or a machine learning group or a group which is working on this our work for us which so it's a question of where you are and what you want to do and what type of organization you are in at the moment. All right.
bottari. Every area is interesting in the future. niche niche econ that question is not really. In the sense that question doesn't have a direct or a simple answer. I can say do Java similar answer if you're expecting it doesn't exist. But all the areas where machine learning is involved is important. Similarly, different person, if you choose to say what about web development, there is a lot of work going with TypeScript. Okay, what about information security, a lot of work is going on there. So he In major spheres of activity, there are a lot of work going on lot of work at different levels going on. In fact, that's something I want to stress again, somebody has again I was out of the job opportunity and a lot of this due to Can we see jobs? The I want to explain for a minute the genesis of this program, this program started about three years back nearly three years now stories are right. Yes, yes, about three years back because the triplicity has such a close connection with the industry in and around the Hydra. So, their senior level people, you are talking to AAA the saying you are producing award 60 masters students I want about 200 graduate, undergrad students and most of them go here to do research or work in us We need people at every level. So can you provide a program which increases the ML knowledge in Hyderabad. So this is the reason for this. programs designed by interpolating when they look for implementation partner since we were also incubated inside ripple it they came to us. But the program is has arisen out of specific expressed interest in the industry for people who can work with a and ml at every level. That's why this program again open door we don't talk about any specific problem domain. See, for example, the digital health program talks about how to use things specifically in the digital realm area. This is a general program which is intended to help you understand how problems in different domains can be addressed as machine learning problems. Or put another way how machine learning a algorithms can be brought to bear to solve problems in different domains.
Yeah, this builds athletes who can be a play any games and How to Become specialists specialists in cricket football, tennis. What do you want to be right? Yeah, this is to get the technical analyzer
blockchain is used when you say trust it's not trust is about some think of it as in a very, very loose waves because as soon as Ronnie will kill me, if you quote me on this, think of it as ATA the authorization authentication, you identify yourself prosystem it tells based on you who you are what you can be allowed to do in a system. Think of when we say trust, think of that in a larger sense. How do we ensure that in a larger systems so that's the type of thing blockchain tries to address.
you'll see for mechanical researchers, if you see an area where there is a large volume of data and you see potentially analyzing the data can lead to insights into problems in your field. Yes, this person Grant will be very useful. side, part time PhD doesn't exist part time. Ms. Yes.
Okay, so there are some questions about the format of the program and also what we do so broadly, I think we should just cover the whole thing. So the program is of four modules module zero being foundations. Seven, would you take through module one?
Yeah, it's module zero is in covers for the Python programming talks about the Python year tool, chain and ecosystem required for a ml. It also covers a little bit of the math required. module one talks about how to formulate problems module two talks about what are the major class of algorithms and a little more detail. Module Three talks about deep learning in detail. But from a very early stage, all the modern topics are introduced, so we actually discuss neural network In Module zero itself, each topic is revisited multiple times. So, the structure is one of what we call spiral methodology where each topic is revisited multiple times in greater and greater detail. You start by understanding at the highest level the techniques, how to play how to formulate a problem to do that. And then you start understanding how to do the algorithms apply these specific libraries first problems, and then you look at how to deploy the solutions as well as understand deep learning techniques. That's how the whole program is okay.
Harish has asked a very important question, is it only a we can program is it a recorded or live online class? So I'm sure ashokan would like to take this
Yeah, it's a live online completely. There are videos available which are required viewing before you come to the class that we have to prepare and depending on your background, and comfort With mathematical obstructions and programming, you will need to spend between five to 10 hours every week in addition to the classes.
More mandatory things to do would be review the
program review or programming. If you have no pipe, if you don't know by then maybe pick up a little Python or attend the bridge course. And most importantly, shore up your abstraction skills. Analysis of forces Dr. Satish Can you say more? I am not sure. I don't know how to interpret that question, Richard. And that person teaches us an analysis
Okay, talking about from a mechanical engineering side. Are you talking about defense forces? It could be anything because I don't know where you're coming from. So if you could add some more I would love it and see whether I can answer the question.
Okay, so there is no question about fees and the program and stuff like that. Let me just take some of those questions. Yeah. So this program, I think we are now running the 15th cohort, right 14 cohorts 14 cohort is going through. In fact, the 13th cohort got probably today's their graduation rate last day. They're completing that program.
this program will be delivered. Thanks to the current situation that we are in and thanks to our platform, which is pretty robust and being used by large institutions. It will be delivered through interactive instructor led live online classes. Right. It was classroom core earlier, but now I don't think the whole calendar year of 2020 classes, people have hundred 200 people coming together is going to be a thing. And this is probably the most important time where working professionals need to ensure their career because of economic downturn, and hence we want to continue with the program and we will deliver it online We have delivered the program almost completely online already one round now, because module two and three of the 13th cohort and the module one of 14 cohort have already gone online and the feedback has been excellent. I think we probably got better NPS than what we used to get in the classroom. Right? Because it was not that much more flexible and much more engaging for participants. The program fee is two lakhs, but I think there are some lockdown scholarships available for people which Susan and other members the team you're not given Swati can give feedback on and can help you with the program. The program starts in the month of August, exact dates will be communicated. The program will be delivered on Saturdays and Sundays it will be a combination of 50% of lectures and 50% of practical work that you will do on your own. All the things you will do on the digital platform and we We'll go ahead with mentor support even outside the classroom so that the mentors are available for help for you as and when you need when you're working on the problem. There are specific officers which are mentioned for it. As a program, you would also be able to interact with triple it faculty and ask them the questions during the live sessions or through forums and adult clearing sessions and recap sessions. So it will it will be as effective or in fact more effective than what you would do in a classroom but now proceed from there. The outcome would be of course certification from triple IIT Hyderabad. And you also become the executive alumni of the AAA at Hyderabad alumni.
Sorry, okay. So that's what it is. And it is, I think, I think any other question about the program Aug two, which meant this will be around for months. Since last August, September, October, November, right. The program used to be three months and it is in the classroom but to ensure that there is a space learning in the online world, there are some additional weeks that are added. The program is jointly offered by triple it and talentsprint This is the only way triple it offers program. We are partners, right? We are partners. The program happens in the talentsprint platform and delivered by to fly the faculty and certification fight or flight faculty. And as much as triple it I think we also have seven other institutions which I was talking to you earlier in the introduction, though as I explained, but them the program is a two lakh rupee fee, but there are special scholarships available.
Let me go back to this question in the meantime, yeah, yes, it's up again. If you have one could argue for example autonomous navigation is a mechanical engineering.
Automotive being a subset
if you are willing to look at it that way Oh, yes. But that's more of a specious answer then a serious answer to your question, but that's not how I want to play any problem where there is significant amount of data available and analyzing that data is going to give us an insight into a systems behavior, the system could be customer response to an app or the system I want to discuss could be customer tendency to stay a customer or dropout or it could be the system the system could be a machine which will wear out a method the word puzzle computational fluid dynamics there somewhere about The other second person has specifically asked the question, I don't know, my friend Professor programs nearby it is a competition. Have you done an expressively specialist? I will catch up with him and ask him but as far as I know, he uses Python a lot, but I don't know whether he uses them Am I know that because whenever we have charted even last week, he knows I am teaching the preliminary components of this ml course he has never discussed anything from it. So if I were to hazard a guess I would say I can't see it being done, but
that's I guess, yeah. Okay. So, there are some question about the effectiveness of online versus digital. So, let me just look at some of this very, step by step. Right in today's context, physical classes cannot happen by any top institution anywhere in the country and the world. Okay. Technology and Education Technology and platforms including diapers dot yay which talentsprint has given replicates and probably brings all the benefits of a physical class and adds to it the benefit of going digital, which is geography. Because earlier this AAA program only people in Hyderabad could join. Now today, I'm in LastPass, we help people from Saudi Arabia, people from us joining the program because it went online. Right. So, in addition to those benefits, it is clearly has effective providing both one on one mentoring group mentoring group labs, group work, and stuff like that. The third part is, I think this is the time at this point of time, most companies are looking at automating faster either for improving efficiency or reducing cost or for both. So that being the case the need for the expertise is no one here and just because the classrooms are not open for one year Most working professional cannot postpone this program by one year right saying that I will do it after a year right. So, it is in this context that the ideal solution available both for the institution which is stipulated here and talentsprint and for the participants is to deliver it in the digital way and I am extremely confident and based on the feedback that I have given out both in batch 13 and batch 14 batch 13 partly went online and batch 14 fully went online. The feedback has been equal if not better than the earlier programs because of the addition additional efforts and focus that the faculty is giving and personal touches our mentors are giving. So, we will continue doing that. And our objective is to make this program as effective even in the online world, and probably that will become the new normal. We are in the session on new normal and we believe that digital will become the new normal and it Including deep tech expertise building in AML. This would be the new normal going forward. ashokan anything you want to add?
No, you're pretty well, but what I see is the
people actually found it easier to even collaborate on hackathons and coding projects on the ML projects, which people are doing even more than the sessions. In fact, people earlier used to find these sessions difficult. Now, they said, the average response was, Oh, these sessions are very much more enjoyable because people could have little more flexibility at what time they join what they work on and things like that in a remote mode. So I would definitely think it is being this.
Okay. There is a question from Pawan. Would you like to take it in the chat?
Yeah. Before that, there is one more question I wanted to Yeah, hurry she have a one to one minute. Free support available which you block
you book a 15 minute or a 30 minute slot and you specify this is the problem you want to discuss, you would like to do it this time, but this time some mentor will be available to talk to you on that. Hands on practice for real time issues. By real time, I presume you're not talking of some 300 millisecond response, but you are generally saying real life
Yeah, so yeah.
Most of the problems are, we see problems which are being solved at some at some point of time. But also remember this program is about teaching you to be a practitioner. So we have to balance the realistic ness of the problem with the pedagogical effectiveness of throwing you into a problem which is realistic, but you can't finish and learn anything is not useful. Giving you a problem, which is a pure textbook problem is also not useful. So the program has to walk the middle path. And in our experience based on the feedback, we have successfully walking that middle.
Thanks so can I think the plan for one Why not? We are already one hour 15 minutes I'm on video. Maybe you can come on video. I think if there is no more questions or maybe I take one more question. Okay. Kumar, do you have anything to ask? Your astronomer and I have been? I have asked you to unmute mute. If there's any question you can ask. No, it
is fine. I already asked
earlier. ticket. Thanks. Good morning. Thanks. Thanks for all your time on a Saturday morning. Good. Thank you very much. Class this was useful and our team will be in touch I think they will also share their phone numbers. really look forward to seeing all of you on the class which starts in August.
All the best for your career in
young people's stay safe.
Watch the entire interview here https://www.youtube.com/embed/1q9luYhoxo4