Return to site

83. Balancing Technology and Research for AI Leaders with Nitin

· podcast,AI


broken image


Podcast with Nitin Singh Part 4


Podcast Title: "BalancingTechnology and Research for AI Leaders: The Pursuit of Happiness"

Summary: In this podcast episode,Andrew Liew interviews Nitin Singh, an experienced AI leader and data
scientist. They discuss the challenges of balancing technology and research,
the future of work, and valuable tips for aspiring data scientists and business
leaders. Nitin shares insights on the importance of having a well-defined
solution design in AI and emphasizes the significance of finding people who can
compensate for your lack of knowledge.

In this podcast episode, Nitin Singh,an accomplished AI leader and data scientist, engages in an insightful
conversation with Andrew Liew. They delve into the dynamic world of AI and its
implications for future work environments. Nitin emphasizes the importance of a
hybrid work culture, combining office and remote work, and how VR technology
may enhance remote interactions. For aspiring data scientists and business
leaders, Nitin advises maintaining continuity in learning, discovering passion,
and cultivating a strong network of peers. Additionally, he introduces his
unique "Path to Solution" framework to evaluate a candidate's
problem-solving abilities during interviews. Towards the end, Nitin shares his
dilemma of balancing technology and research while seeking advice from
listeners on finding happiness and meaning in his career.

[00:00:00] Andrew Liew: Do you think that the communityout there those listening to this podcast, we should actually come together to
create some kind of like, the open AI, like open source AI safeguard, a chapter
where not just the application of AI, but also the mechanics to, create a ring
fence around this technology or the use of application so that the

[00:00:20] Andrew Liew: These days I have a few friendswho went for interview and the people I insisted, they really fly over to see
the guy face to face. It is a real skin. It's a real gut. No, especially the
last round before the signing it's happening.

[00:00:34] Nitin Singh: See, my personal opinion is theonly way we can have a regulation is people like Sundar Pichai, all these open

[00:00:43] Nitin Singh: See, you have to be very carefulwhen you listen to. I think the OpenAI who was that guy? I forgot his name,

[00:00:49] Andrew Liew: I know who you're talking about.Yeah.

[00:00:51] Nitin Singh: One second. He was on the worldtour, yeah, I forgot. How can I forget, man?

[00:00:54] Andrew Liew: No worries. These days there's somany information. Ah, Sam Altman. Yeah, Sam Altman,

[00:00:59] Nitin Singh: yeah.

[00:00:59] Nitin Singh: [00:01:00]You listen to Sundar Pichai and Sam Altman. Even Sam Altman said that the
future would not be this large, heavy models. And these people are the first
one to know how, far the technology can go. They can see that. And that is
where it's very important for these people. Let's say Elon Musk Sundar Pichai,
all these big companies, where a lot of research is taking place, should
actually come together.

[00:01:24] Nitin Singh: And they can only come togetherif they have a business objective in place. And it's very difficult for them to
come together because the normalization which takes place after they join hands
may impact any company adversely, may impact some company positively, right? So
that's the thing. But unless they do it in collaboration with the government, I
think it's not going to happen.

[00:01:44] Nitin Singh: It would be very difficult for usto probably create a community, talk about governance, but ultimately the bucks
are in their hand.

[00:01:51] Andrew Liew: So you don't think it's like thestory of this a boy who told a starfish, right? There was this. Tourists that
look at this little boy, Why are you throwing this [00:02:00]starfish?

[00:02:00] Andrew Liew: There's like hundreds andthousands of starfish along the coast. Can you say, Yes, I can't get the
government to create a machine to but I can just do it one by one, one by one.
You don't think at the individual level people like you and me data scientists
or leaders, Can actually do the little parts.

[00:02:18] Andrew Liew: Like for you in your case, maybecreate this ethical box in, whoever transacting on Molodin or something like
that. Don't you think so? I don't.

[00:02:26] Nitin Singh: I will tell you, see, in ourlife, see people like us who are coming from a, let's say middle class or low
middle class family for them.

[00:02:35] Nitin Singh: The first purpose is to actuallymanage their family, right? Yes, I agree. All this any, what do you call that?
Initiative at a high level are driven by people who have money, right? That's
my feeling. Okay, I feel that for people like us, it's very difficult to think
beyond our families because we are not at that level where we can stop working and
our family would be fine.

[00:02:56] Nitin Singh: Now, what I believe in, karma,actually when you look at a Hindu [00:03:00]mythology, it has some duties. So as per Hindu mythology, the first quarter of
your life, let's say first 25 years, you focus on learning next 25, you focus
on remain managing a family taking care of them.

[00:03:13] Nitin Singh: Third, focusing on then folkmeditation, God so they have divided life in such a way. And if you believe in
karma, as in what you do is It defines you. So for us to for anybody who is
born, the first task for them is to actually take care of their family. If
you're not taking care of your parents, your family, and you're trying to do
something else, I think that's, a little contradictory because he's right.

[00:03:35] Nitin Singh: So you have to take care of yourfamily first. When they're fine, you try to learn. Then you start looking to do
something with community. People like who are at that level can actually make a
change. Now people who have actually come to that level. And also influenced by
the business side of it, right?

[00:03:51] Nitin Singh: So the ethics, forget about AIethics, forget about the human ethics, that's very important. So people who
have a good moral fiber [00:04:00] good, ethicsin themselves, they would be the right people to lead these kinds of
initiatives. Okay. Because they would be focusing on the real problem, they
would be fighting really hard.

[00:04:08] Nitin Singh: But the irony is it's verydifficult to find those people in bunch the very less people like that. And
even though. When you talk about me, I don't think even I am that much, Grown
that I can do that thing. So very honest. I know my capabilities What I can do
what I cannot do.

[00:04:23] Nitin Singh: I can only make people learnthings. Okay, that's

[00:04:25] Andrew Liew: What i'm saying, maybe you caneven create just I call it like guardrails for your e commerce site Like to
make sure that like I give you an example like just this few weeks. There was
the police actually tried to Catch this and that guy actually, I think Jill few
years ago and he was selling fake goods on carousel.

[00:04:45] Andrew Liew: And then a few years later whenhe was released from prison, he still got caught selling fake goods. So the public
was what is going on? Why isn't carousel creating authentication of. Is this
guy gonna sell the same type of fake goods? Because of course, we have to give [00:05:00] people a second chance. We have to givegrace because it's service time.

[00:05:02] Andrew Liew: But we have to stop bad actorsfrom repeatedly sell the same fake goods.

[00:05:07] Nitin Singh: You're correct. You're correct.There are actually two different things. I was talking from a very high level
perspective of companies coming up with the models who have different
capabilities. But I think at a company level, When you're creating a product,
it's very important to have those guardrails.

[00:05:20] Nitin Singh: I call it a rule based layer.When I say that if a model is 90% accurate, you need to put that 10% in form of
rules or in form in some other form that you can compensate for that. I think
guardrails are very important. And even I feel in carousel for that matter, a
lot of things can be AI based.

[00:05:36] Nitin Singh: So that it's, see for me also,it's very difficult to be honest, right? I, would totally agree with the people
who have bought fake goods via carousel. So I am also one of them. Yeah.

[00:05:45] Andrew Liew: So, you bought like that, thatyou got a shop that it's, fake or it's beyond your expectation on the price
that you pay.

[00:05:53] Nitin Singh: I got a broken product actually.I got a camera. It's for the camera for my my son's room. So my, I have a three
year old [00:06:00] son. Okay. And if I put abig camera there, he would go and plug out the camera because a lot of time
when he gets up in the night, I do camera. I tell him I get a notification.

[00:06:08] Nitin Singh: I tell him to go to the bed threetimes. He understood. He understood. So he just take the plug out of the big
cameras. I was thinking if I can put that camera somewhere hidden so that he
doesn't, he cannot go and so that would have ordered. I got a broken one. So
anyways, so I agree. I think those guardrails are very much required in your

[00:06:25] Nitin Singh: Let it be e commerce or financeor anyone. I think in public facing product, it's very important. People should
not be scammed. I think that should be the highest priority. You're correct.

[00:06:34] Andrew Liew: Okay, let's come back to anotherinteresting part about what do you think is the future of work?

[00:06:38] Andrew Liew: What will work look like in thenext five to 10 years? It will be

[00:06:42] Nitin Singh: hybrid. I feel so. See, as amanager, another individual contributor as well. I find office changes. People
need some change out of their life, right? The whole time you're sitting in at
home. So sitting at home was good initially because you used to go office every

[00:06:58] Nitin Singh: It was good initially, [00:07:00] but then you would realize the kind ofefficiency you have is reduced to a bit. So I have noticed that. So I think
hybrid culture is something which is, I think, is a future where work from home
would become more interactive through VR or maybe through.

[00:07:14] Andrew Liew: What do you have to say to thoselike who are a bit old fashioned or still cling on to I want everybody back to
the office, RTO all the time.

[00:07:22] Andrew Liew: I want to see them FaceTime.What's your view on

[00:07:24] Nitin Singh: it? It's not going to happenbecause, see, generation has changed. Now when I look back, people from my
generation, I'm the manager team and now when I'm managing people, nowadays
focus a lot more on their work life balance. This was not the case 20 years

[00:07:40] Nitin Singh: It was easy for me to get thework done. Call them. Okay, let's come on the weekend. But I would also ensure
that I'm also there. So I would not never just ask my team to come and work. I
would be there with them. But later on in my career, I realized when I asked
somebody, Hey, can you come on the weekend?

[00:07:53] Nitin Singh: He would say, no, I have to go myhome this weekend. Then I say, okay, next weekend, what's the plan? Okay, next
weekend also I'm going. And they'd be like [00:08:00]just so clueless. They pretend to be clueless that Yeah, I have to go home. I
cannot come. So, that's the difference. So you have to manage those people
accordingly, plan your project accordingly, manage your risk in advance so that
you don't have to extend any time.

[00:08:13] Nitin Singh: So that's how it should be. Butyeah, the people are different. And that is the reason you cannot manage the
new wave of talent coming in with old way of management. It is very hybrid
only. I don't see any reason why. And in fact, people would start going to
office not for nine hours, maybe four hours, five hours, get the above them
being more efficient.

[00:08:32] Nitin Singh: And then come back. I think thatis the flexibility I'm seeing in the coming time. So

[00:08:36] Andrew Liew: you think that the future ishybrid, right? So in other words do you think that like on site is more about
interaction, learning of ideas and so called maybe even verifying works,
whereas like those online meetings, it's just to have a quick checklist stand

[00:08:50] Andrew Liew: Okay. Everything. Okay, good.Just keep going. Do you think that will be the future or and especially Those
workplace technology like styling Google styling, where, when I see you face [00:09:00] to face, it's almost like real

[00:09:01] Nitin Singh: what do you think? I think that'smore of a, selling point.

[00:09:05] Nitin Singh: I think what importance is it'snot that I can see somebody like real life, I'll be more comfortable talking. That's
not the case. Okay, it's and it's very standard that whenever you meet
somebody, whenever you have a meeting, you need to have an agenda. If you don't
have an agenda, it doesn't matter whether meeting face to face on call, you
need to have an agenda.

[00:09:22] Nitin Singh: And that is where scrum comesinto the place. You need to have a defined point, but you want to discuss and
that's about it. See, the real discussion takes about 5 10 minutes only in my
calls with client the, office work, the business work is only for 10 minutes.
The rest 20 minutes we're spending just chatting, getting to know the other

[00:09:37] Nitin Singh: So I don't think it's verynecessary for to meet people in in, person, mainly also because when you go to
the office, everybody have their schedule. Even in office, you would meet them
for half an hour for that slot of meeting. It's not that in office you're
meeting that person 10 times and you're getting to know more.

[00:09:54] Nitin Singh: It's not like that. I think it ishybrid, but you need to make sure at least calls [00:10:00]on call to be more efficient. I think same holds with the in person call as
well. When you're talking to somebody, you need to have an agenda. What you're
planning to get out. What is the outcome of the whole exercise? I think that's
the process we need to have.

[00:10:11] Nitin Singh: Like I said, the process is veryimportant. If you have a process in place, things would be fine.

[00:10:15] Andrew Liew: I see. Okay. Sounds good. Okay.For those who are like fresh graduates or people who work with two, three years
experience and they really look to inspire to people like you what is your tips
for them to grow their career as, a data scientist leader or AI leader?

[00:10:31] Andrew Liew: What are some of the tips you cangive to them?

[00:10:33] Nitin Singh: Not only data science for anyfield. I think we need to be patient. Okay, we need to have that patient
because right now the lot of resources online and you cannot learn everything.
So you need to maintain continuity and go step by step, right? Try to solve the

[00:10:50] Nitin Singh: Understand the business problem.How did you solve it? Then the top down approach I was talking about. Follow
the top down approach. Being continuous in your effort. I think these are the
only two things one should be [00:11:00] therewhen one should do passion is very important, but I don't know because I was
very passionate about data sciences to be no data science to be honest, it's.

[00:11:08] Nitin Singh: To differentiate my work. I thinkI was passionate about that. So everybody should have that trigger, but what
their passion is, it not necessarily could be data science. It could be
somebody, something else. So if they have their passion in themselves, that we
want to achieve that goal. They're really passionate about it, then these steps
would really help them grow faster.

[00:11:28] Nitin Singh: Otherwise, it's very difficult. Idon't know if you're not passionate about something, and that field requires a
lot of effort, like probability, optimization mathematics, linear algebra
probability, a lot of things are there it's not just one field. And then at the
top of it, you need a good communication and a business sense as well.

[00:11:45] Nitin Singh: So it's everything, man. What isleft? So that's, I would say to new people coming into the field, find what
their passion is. Even if they don't find their passion, just maintain
continuity in whatever they study. Solve problems, be [00:12:00]continuous, and your average of six people you meet, like I said.

[00:12:03] Nitin Singh: Have a good friend circle.

[00:12:04] Andrew Liew: Like you say, so let me summarizea bit, because you have so many points of nuggets. So one of the key things is
maintain continuity, it's a continuous team. Find your passion. You, are some
of your six friends, now coming back to the next interesting question is what
about those business leaders and let's say they want to build a data science
capability or AI capability.

[00:12:25] Andrew Liew: What are some of the things thatyou think they should really think about, especially from scratch?

[00:12:28] Nitin Singh: See the business alreadyestablished business leaders. I think they would have a better sense of
business than anybody else. Agree. So at that level, if let's say I'm in their
shoes, so I would actually look for people who actually compensate for my lack
of knowledge, right?

[00:12:43] Nitin Singh: So I will look for those kind ofpeople and that's how it should be. I also try to hire people who are more who
knows more than me. So that I learn more. It's not the other way around. So for
those business leaders, you need to have that kind of team where they can then
take a business decision. Now, you don't [00:13:00]necessarily need to know the technical aspect of it.

[00:13:02] Nitin Singh: Because you already have a SME totake care of that area. It's, not as easy. Like I said, it's very difficult to
find that person and the synergy between two people. And that is one of the
reason I couldn't find that synergy. That is one of the reason I also try to
foray more into the technical part as well.

[00:13:18] Nitin Singh: Yeah. But that's important. Youneed to have a team. It's, always at a team at a business level where you find
people who are more talented than you. Who compensate your lack of knowledge in
a domain and then you take a decision. I think that's and follow the pilot man.
If you are rolling out data products, follow the simple rule of creating a
hypothesis pilot, check the outcome and then roll it out.

[00:13:40] Nitin Singh: I think that's what it should betalking

[00:13:42] Andrew Liew: about, finding people who arebetter than you. What are some of the tips that you can actually even advise
the business leader, even some of the data science director manager, to find
somebody better than them so that both parties can improve? What are some of
your tips for them?

[00:13:57] Nitin Singh: I think it's very difficult toright now [00:14:00] hire people, especially indata sciences, mainly because, All the same.

[00:14:03] Andrew Liew: I thought SQL, RDS data sethypothesis, all the keywords, right? All the keywords.

[00:14:09] Nitin Singh: XGBoost. Just, hold on. Sure, Youwill find XGBoost generative

[00:14:13] Nitin Singh: AI. A lot of people usegenerative AI in their headings in LinkedIn, you'll see. So things change very
fast. But yeah just hold on. I think yeah, I forgot the question, man. What
were you asking?

[00:14:24] Andrew Liew: How do you know, how to findsomebody? Try to hire Oh, people are better than you.

[00:14:30] Andrew Liew: Because like you say, on paper,everybody have the same keywords. What are some of the things that only when
you the tips that you can give to leaders to Ah, this guy is somebody that I
want to hire or even based on your personal experience. Yeah.

[00:14:43] Nitin Singh: So I think I believe in aframework where you talk about path to solution.

[00:14:47] Nitin Singh: These are my, personal frameworksare not... Please explain

[00:14:51] Andrew Liew: What does path to solution

[00:14:52] Nitin Singh: framework mean? So when I'msolving a problem I follow a journey. I have a journey to a solution. For [00:15:00] example, there's a product mismatchproblem in car industry as well. People would write a description of a car in
different ways, right?

[00:15:06] Nitin Singh: Somebody would say like Daihatsupickup. Somebody with Daihatsu PU Daihatsu pickup truck. Different way of
writing and it's a free text. It's a product mismatch problem. The way I
approach that problem, I solve that problem. I'll put my the person I'm taking
interview on the same journey.

[00:15:22] Nitin Singh: I give him the same problem, givethe same input I have, and try to nudge him in between that these are the other
options we have available. And if I see, if that person is able to think the
way I did it, at least initially, or think better than I did, I think that is
what you need. You need somebody who has good aptitude, good thought process,
now, creating model is not difficult. Yes, it's about how you design the
problem, how you design the solution. And that is when I say the solution
design is the actual skill data scientist is not the actual skill is how you're
going to do the solution design in AI. The solution design is, I feel. Is the I
would say the profile, which [00:16:00]probably in coming time, you will see come more often.

[00:16:02] Nitin Singh: People will talk about AIsolution design.

[00:16:03] Andrew Liew: When it comes to solution design,you only know when you interact with the guy, right? Because it's more like an
art, right? Because the sciences is like the code packages, the variables.
Anything that you can find on the you can implement them. But the things that
you can't find is, like you said, like in that for example, that car.

[00:16:19] Andrew Liew: What are the features? What arethe considerations like those kind of like, how do you even know

[00:16:25] Nitin Singh: yeah, I think you're making myanswer look better by providing these, context. So yeah, I appreciate that. So
you're correct. You're totally right. Because these things you cannot find on

[00:16:36] Nitin Singh: You cannot find anywhere. So I thinkit's about the case based interview, which I really focus on when I started.
Thank you. I used to ask, okay, what is a bias, variance, trade off? What is p
value? What is cost function? It doesn't matter, man. It's just useless to ask
those questions. The thing should be to talk about a problem and then look at
how they think to solve that problem.

[00:16:57] Nitin Singh: I think that is the skill youneed, because [00:17:00] when you're workingwith that team, with that teammate, you're actually discussing the business
problem. Coming up with a model is... It's not difficult classification problem
or regression problem or a gender tax generation problem, object, whatever the
problem is, we already have schools for that.

[00:17:14] Nitin Singh: We already have models for that,but we don't have something where people can design and think about a solution
in the right direction. It's like a consultant, to be honest they have the
framework, how to solve a problem. If a company is not doing sales well, they would
look at the revenue, they would look at the cost, product cost they have the
whole framework to come up with the solutions.

[00:17:34] Nitin Singh: Just like that only. In AI,machine learning, you need to be really good in designing and thinking about
problem in a way that Compliments you and that's how you solve a problem. And
that's how I even I interview people, give them a use case and see what's their
chain of thought, what's the path to solution they have.

[00:17:49] Andrew Liew: Wow. These are very useful ideas.So you know, you have shared so much from your backstory to the future of AI,
to tips on future work and graduates and business leaders. [00:18:00] What do you want to ask from this audiencewith my followers of 16, 000 LinkedIn followers, what do you, for those
listening to this, what do you want to ask for?

[00:18:08] Andrew Liew: What's your universe ask? Yeah,go ahead.

[00:18:10] Nitin Singh: So I have a different, I have adifferent perspective and I, so I always look for ways to be happy. It's very
different. So I'm not going to talk about it here. What's important is I always
look for people who have somehow Practice this art. A lot of times I keep worry
about learning more.

[00:18:28] Nitin Singh: I'm not learning. If in a singleday or a week, I'm not able to learn anything. I get restless. That's not good.
That's not good for your health. So if somebody can tell me how exactly they
manage this kind of, what aggression toward something where you are actually
getting stress not doing that because I'm still trying to figure out the right

[00:18:46] Nitin Singh: I have two kids, I also need togive my time to my family as well. And I cannot do research. I really want to
do research, but I don't have time for that because. If I start doing research,
then I cannot deploy this product for my company. Ultimately, that's [00:19:00] what I'm being paid for. So I'm in adilemma where I have to, do I go into research or do I keep doing what I'm
doing, but I need to do, I need to learn more technical technology need, need
to be more research oriented.

[00:19:13] Nitin Singh: So I'm still trying to battlethat thing. But yeah, I think in Gartner, I had a one. One manager, she, she
told me, yes, you explained me the whole thing and that's where it got a little
better for me. And since then I've been focusing more on a business side
because that is more important.

[00:19:29] Nitin Singh: But yeah, this again, I wouldlike to know how people actually took care of these situations where they have
walked between technology and research, how do they do that? Yes,

[00:19:39] Andrew Liew: Like definitely to the audienceout there Nitin has shared so much views, so much tips.

[00:19:44] Andrew Liew: If you are listening to this andif you have any ideas, view or resources or way of thinking please, Nitin will
share with me the LinkedIn handle or social media handles. You can actually
reach out to him offer your view in terms of your perspective of how to. Can I
say this is like[00:20:00] the, pursuit ofhappiness or the pursuit of meaning?

[00:20:03] Andrew Liew: Is it something like

[00:20:03] Nitin Singh: that? Yeah it, actually isbecause ultimately, down the life, if you see how your life is not about what
you're doing. It's about being happy, being making people happier around you.
And that's what the ultimate life should be. People, you should help people.

[00:20:18] Nitin Singh: Like in Singapore, what I reallyadmire is people actually helping others when you go out. So they'll open the
door and you start doing it. If you're not doing it in your country, you'll
start doing it here. So that's what's more important. People around you should
be happy, comfortable, and that should be your main, objective in your life.

[00:20:34] Nitin Singh: So that's your correct. It's likepursuit of happiness, to be honest.
[00:20:37] Andrew Liew: Okay, great. Thank you so much,Nathan. We have come to the end of the podcast.