Vizuara · Transcript · · Updated
Ground Truth Ep.3 | Prof. Prathosh A.P, IISC Bangalore – Transcript
Video by Vizuara · Watch the original on YouTube
Summary
AI-generated from the transcript. Verify details against the source video.
- Professor Prathosh A.P. completed his PhD at the age of 25 with a specialization in pattern recognition and machine learning for voice models.
- Prof. Prathosh A.P envisions a university model that focuses purely on teaching students how to work with AI agents rather than tying education to monetary value.
- Prof. Prathosh A.P argues that India needs to build both frontier technology models and frugal AI stacks to maintain control over the intelligence and cultural narrative.
Use this video with Pro
Reading, copying and transcript downloads are free. Create saved AI results for this video with a signed-in Pro account ($9/month).
Transcript
Search ignores accents. All transcript text stays visible.
So that I when I grew up I did not know that institutions called IIT existically but one thing which is phenomenally changed today is that intelligence has now stopped being valuable. I think that is where the the biggest changes happen where uh remembering and reproducing information has uh as a skill is not valued anymore.
We make our children to be successful. We never make we never groom them to be happy. Being employed versus starting a business is like being a pet dog versus being a street dog. If you're a street dog, then you'll have to fight for your territory. You'll have to fight for your exist just last week where he said that human beings will become like cows. AI I'm worried today that the idea of
morality idea of what constitutes good or bad is uh is is is coming from YouTube and Instagram reels and shots today. For those of you who are currently working at top companies in the US or Europe in Meta, Amazon, uh, Nvidia, whatever that there are people in India who really want to make things happen. Humanity would collectively with these agents would solve problems like cancer. No,
>> if it does not, I think all this is uh is a waste. All this is only a game is learn two things. Learn a language to communicate and learn a language to think. Hello everyone and welcome to the ground truth AI with Vijara. Today we have professor Prat with us as our guest. Uh
he is one of the leading AI researchers in our country and we are very very excited to speak with him. I'll explain how I came to know about professor. uh two weeks back I was sitting in a cafe in Pune and uh his video came in my Twitter feed and usually when I watch videos uh I don't watch the entire video but usually 2 to 3 minutes but somehow I found the conversation very honest he was just explaining what's coming in his
mind and u the conversation was very genuine and and and honest so that was the time I thought okay it will be very interesting to speak with him and get to know more about him what his thoughts are etc. So, uh we'll be of course talking about a lot of topics today. Uh I'll I'll start out with your journey professor. Uh I noticed a very interesting path in your journey where
you have a good mix of academia and and industry. A lot of our friends ask us hey should I go into academia or industry? For them it's either it's it's it's like binary either academia or industry. in your profile I noticed that you have a very good balance of both and I'm very curious to understand how you have maintained that balance u you have started one company back I think two years back and now you have another company I don't know if you have both
the companies are still active right now so I'm very interested to maybe you can start with that what was your thought process before entering academia was it a deliberate choice to balance academia with industry or it just happened and what what were your motives what do you gain from industry that you don't get lit from academia.
>> Yeah. First of all, thanks for having me. Uh so as I said uh u I mean I've been following you and the good work that you do. Uh I think we are also connected on LinkedIn and so on. So um yeah you asked too many questions.
>> What is my journey and why academia versus industry? It's a uh it's a journey. So how it all happened was I come from a small town which called Mysore which is famous for its uh tourist significance. Um we grew up I don't know about you I grew up in times where we did not have internet.
>> Uh so whatever your parents or your extended family would tell you is is is all that you know. So you wouldn't believe if I tell you that I when I grew up I did not know that institutions called IIT exist.
Uh so our families nobody went to IIT. So we did not know. Uh interestingly when I was in class 12 our goal was to get into an engineering college which is the best one in our city. Nothing outside of our city. Right. So there was going outside of Mysore was was not even an option. Now uh class 12 somebody told me that there is an exam called J means those days it used to be called as Ale
E. So they said you write that exam uh so that will become a uh preparation for the state uh engineering entrance exam. So our goal was state engineering entrance exam but you write J means because it will be a good preparation for so I never went to any coaching centers etc. So I came to Bangalore to write J means turned out that I got a three-digit rank uh in in J that was a preparation exam. So despite that my
mother told me who would go to NIT it's too far away to stay back so we have to be in our uh in our city. So then I ended up joining one of the colleges in local colleges in Mysore. So it's called JC one of the top ones in Karnataka at that time. So I joined there. Um from my childhood I wanted to become a teacher because I was mentored by a lot of good teachers. So I wanted to become
a teacher. Okay. So I joined engineering. Uh then I I met very good mentors. You know one thing if I am very thankful about uh thankful in my life about is the kind of mentorship that I received. See that's why I think that uh that even though you have so many AI tools etc. the correct kind of mentorship that you get at right points in time uh in your life matters a lot.
So fortunately there was one mathematician uh who was in TFR. For some reasons he quit his TFR job and came and joined our college as a math prof. This is professor Yoga Nandanda.
He mentored uh IMO team for India. This is a pretty good number theorist and mathematician. He came and that changed a lot. I mean he was the one who actually first told me that an institute called IC exists >> or IT exists and he taught us math for all eight semesters. So I fell in love with mathematics and the engineering applications of it and so on. Then one thing led to the other. I went to I was a KBPY fellow uh that time and he
encouraged me to apply to all of them and started doing stuff and then my uh liking towards teaching changed towards becoming a teacher in a in engineering in an engineering college.
>> So then I asked my teachers what does it take to become a teacher in an engineering college? I said you need to get a PhD. I had no idea what a PhD was. Okay. So then okay then I said okay my the ultimate goal is to become a teacher. So now uh get a PhD. Now again going out of country was not an option because of multiple family reasons etc.
So I thought okay what's the best college in the country? It was IA is what people told me. So I approached a few faculty here and then they said okay there is a program called direct PhD. I think a lot of people does not know this. In a lot of good colleges across the world, including the Stanfords and Mats of the world, you will be you can be a student can be admitted for PhD right after bachelor's.
>> Mhm. >> Okay. A lot of people think that it's after masters. I'm sure you people know it but a lot of people do not know this. >> Sure. >> Even in Indian universities like IITs and IC, one can join for a PhD program right after bachelor's. >> So my advisor told me that if you're anywhere clear about PhD, why not join right after bachelors?
So this was in which year? 2019. >> So then I asked him I didn't know that it was a possibility. Then I understood that it was a he then I asked him how does one get into a PhD program. Then he said there is an exam called GATE you'll have to crack that and then interview whatever process was that I just did all that meticulously and then I joined for the PhD program here in IS itself. So then again for lucky and fortunate to
have very good mentors. Uh I could finish my PhD in 3 years. So um >> I was a PhD before I I mean I I was 25 years old. >> Oh it's probably some kind of record right? >> I don't know.
>> Apparently not. There was one of my friends who submitted little earlier but the point is yeah so I could do it earlier. >> What was the topic of your PhD? >> I worked on uh pattern recognition ML for speech for what what what is called as voice models today. We were building some of that.
>> Oh okay. Okay. >> Yeah. >> So my PhD was on ML for voice. Okay. And again got extraordinary mentors here and I was lucky that all my papers got published. Uh it is very unlikely but it just happened. So then comes the industry part. U before I could submit the thesis I just wanted to uh see what industry is and one of the labs here which is Xerox research. They had just set up a new research lab here. The polar auto
research center had set up a lab here and they were looking for interns and my adviser had a collaborator there. I went there for an internship. >> This is PhD internship. >> So then uh they >> during the PhD you did an internship.
That's it's very rare, right? >> No, it's not. I mean that is one thing that people don't know. Uh good labs would want PhD interns. In fact, now in my lab as a faculty, I've sort of mandated that all my students have to intern in a good lab for at least 6 months. All my PhD students have interned in places like Google Deep Mind, Microsoft Research, Adobe Research, IBM and so on. I I ensure that they go and look at the industry because see I'm of this belief that if you're an
engineer, if you're not building things, uh then you cannot be a researcher in engineering. >> So every engineering researcher is ultimately a builder, right? Something has to be built. So you have you need that industry exposure. So then I went to internship there and then one of my mentors there again told me that um while the target is here you are shooting there that's exactly what he told me because most of academic problems do not get translated into uh
real stuff is what he told me that was a bit extreme but still then they made me an offer again the goal was still to become a faculty in in in go back to mysore and become a faculty right never the farthest that I went to mysur was Bangalore which is 150 kilometers wanted to go back then to be a faculty you need at least a good two two to three years of post-doal experience then I said anyway these people are making me a full-time offer let me stay that was a
research lab >> so stay there for a couple of years and then start making applications uh spent 2 three years there again met fantastic people uh the one of my managers was is the current director of Google deep mind in India >> I was mentored by very good people this is Dr. Manish Gupta. So mentored by great people. Then I realized that uh I cannot go back to my undergrad college.