YT YouTube2Text

YouTube2Text → Transcripts → The Art of Building a 2,400-Person Organisation In 21 Months Ft. Aditya Khera, BMW TechWorks India

The Art of Building a 2,400-Person Organisation In 21 Months Ft. Aditya Khera, BMW TechWorks India — Transcript

by ETGCCWorld · 5,412 words · 935 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:09So very often the trick, you know, if
0:11you are sort of managing a team,
0:12managing a function, managing an
0:14organization,
0:15uh the trick is never to have all the
0:19right answers.
0:20The trick is
0:22do you ask the right questions to really
0:24understand
0:26and break that complexity of a problem
0:30into strategic chunks, into business
0:33levers. When you have product ownership,
0:35you're not waiting for instructions.
0:38You're not waiting for requirements.
0:40You define the requirements, so you have
0:42shifted left, you are upstream now.
0:43Because you own the product, you define
0:46the requirements, and then you build it,
0:47release it. You know, for decades we
0:49have loved our car.
0:51It's time for cars to love you back,
0:54right? So, which can only happen with
0:56intelligence and and software. And I
0:58think one of the biggest promises why
1:00software is such a game-changer is that
1:02it keeps your car current, right? It
1:04doesn't hook you to model year.
1:16>> Most GCC leaders inherit an
1:18organization.
1:20Aditya Khera
1:21inherited an idea. In 2024, BMW Group
1:25and Tata Technologies came together to
1:27create BMW Tech Works India, a new
1:31technology organization that would work
1:33across automotive, software, IT, and
1:37engineering. It started with about 100
1:40odd people, and within months, it it
1:42crossed 1,000.
1:45But the more interesting story isn't how
1:48quickly it hired.
1:49It It's what it was being built to do.
1:53Today, I have the person building that
1:55capability in India, and the person who
1:58has had an unusual journey himself
2:00across functions, businesses, and
2:03geographies. Aditya, welcome to GCC
2:05Unplugged.
2:06>> Thank you, Sourav.
2:07>> So, as I said that you had a very
2:09unusual journey, Aditya. You know, you
2:11moved across functions, businesses, and
2:14geographies rather than
2:15you know, building your career around
2:17one specialization. So,
2:20was that deliberate or you know, did
2:22your career only start making sense
2:25you know, when you look at it you know,
2:27from
2:29look at it backwards?
2:31>> I think it's a bit of both.
2:33So,
2:34when I took those choices, they were
2:36deliberate, but I didn't know how the
2:39dots would connect. So, it looks uh
2:44it looks like it sort of fell in place
2:46so far, but I I didn't think I had a
2:48grand plan. So, my motto was very clear.
2:52Uh when the learning curve was flat and
2:54I thought I'd given it all to the
2:55organization, I wasn't adding you know,
2:58more value than I could possibly and
3:01so, you know, there was a time when you
3:03know, you could take a plunge into
3:04another opportunity and since
3:07I
3:07I was fortunate to be part of Tata
3:09Administrative Services
3:11where there is actually the system
3:15nudges you to experiment in the first
3:17few years
3:19and then you anchor down.
3:20So, it was a bit of both.
3:22>> Okay. All right. So, since you've
3:24mentioned Tata Administrative Services,
3:26what did the
3:27what did the experience really
3:30teach you about leadership that you
3:33know, you would not have learned by
3:34probably becoming a
3:36specialist maybe?
3:38>> I think you know, when we were
3:40in TAS, at least for me, one of the
3:43bigger fears was
3:46that you
3:47enter into a company
3:50at somewhat inflated grade
3:53and you have no depth
3:56and you're supposed to lead a team which
3:59has been in the industry for you know, a
4:02decade or two.
4:03And it was very intimidating.
4:06So, I think there are few things that I
4:07learned.
4:08One is
4:10respecting people who have domain
4:13knowledge, right? Because you can never
4:15catch up to that, right? So, you have to
4:18defer to them on certain decisions,
4:21certain issues. Uh
4:24equally, uh
4:25when you tend to move in the early
4:27stages,
4:30you quickly get into this knack of
4:33really
4:34distilling the core of the issue you're
4:36dealing with. You
4:38learn to ask the right questions.
4:42So, very often the trick, you know, if
4:43you are sort of managing a team,
4:45managing a function, managing an
4:46organization,
4:48uh the trick is never to have all the
4:51right answers.
4:53The trick is
4:55do you ask the right questions to really
4:57understand
4:59and break that complexity of a problem
5:03into strategic chunks, into business
5:05levers
5:06that you then monitor
5:08by depending on the same specialist, by
5:10hiring, you know, team members who can
5:12commit to the cause.
5:14So, if I were to distill back,
5:17I think this
5:19movement from
5:20function to another function, region,
5:24industry, I think it gives you a lot of
5:26perspective and it it gives you a range
5:30and you can start connecting the dots
5:33that maybe somebody deep into a field
5:36may fail to connect. Yeah, so that gives
5:39you a vantage point, so it's an
5:41advantage.
5:43Uh
5:45I never consider it a disadvantage
5:47because I don't think I can ever catch
5:49up to a specialist. So, my
5:52style is to ask the right questions,
5:54understand, help them understand where
5:57strategically the bottlenecks are and
5:58help them solve it. And I think that's
6:01how you develop mutual sort of trust and
6:03respect. So,
6:04never doubting your domain specialist,
6:06always respecting them.
6:08Uh getting a perspective because you
6:09have
6:10seen several situations in several
6:11industries, so you can connect the dot
6:13connect the dots and sort of arrive at a
6:14pattern that somebody deep in a tunnel
6:16may not be able to uh and then just ask
6:19the right questions, distill complexity
6:21of a problem into chunks and business
6:23levers that you monitor uh and let the
6:26team of expert specialist do their job,
6:29give them the autonomy and the respect
6:31they they deserve.
6:32>> So, talking about specialist uh
6:34Aditya and you know, today technology
6:36increasingly rewards uh
6:39you know, deep specialist, be it AI
6:41engineers, cybersecurity experts,
6:44uh software architects, and all.
6:47Yet, you know, your career is almost an
6:50you know, sort of an uh
6:52argument for a generalist. So, do you
6:54think in you know, this age and time,
6:57we've started uh undervaluing a person
7:01who used to be called a general manager?
7:04>> So, one of my ex bosses used to say
7:08um
7:10that eventually you need to be a
7:11T-shaped professional.
7:12>> Yeah.
7:13>> Right?
7:14So, the horizontal part of the T comes
7:16with lateral movement that happened for
7:18me.
7:19But at some point of time, you need to
7:20anchor down in an industry
7:22and have some roots. Now,
7:25if I talk about myself,
7:27um I went from
7:29a refrigeration company to a chemicals
7:32commodity company to Bombay House.
7:35Uh
7:35and then
7:37I decided to anchor down in digital and
7:39engineering services. So, that's been my
7:41anchor from 2016 onwards, right?
7:44Now,
7:45the point I'm trying to make is that
7:49there is there is
7:51little merit if in being only a
7:54generalist.
7:55You need to have some anchor point.
7:57Right? So, some of my peers who moved
7:59companies,
8:00they
8:02created a specialization in a function.
8:04So, they may have started with marketing
8:06for a salt, then they moved to tea to a
8:09commodity but they maintained a core.
8:12Uh for me the core was strategy sales
8:15and business development. So I
8:16maintained my core while I shifted
8:17industries.
8:19And then from an industry perspective
8:20itself I anchored down.
8:22So
8:24if you have an anchor point
8:27to me generalist can actually create a
8:31lot of impact. Going back to the first
8:34response I made that you know you just
8:35need to understand the business.
8:37Uh if you have a core and you have seen
8:39the world you can connect the dots
8:41faster and you can sort of make things
8:42move.
8:44Uh
8:45I I know where you're getting from
8:47because uh
8:48the industry in general is you know
8:50maybe getting more nerdy and geeky. Uh
8:53but I think at the end of the day
8:56if you're running a big function or a
8:57company you need an orchestrator.
9:00And by definition if you're
9:01orchestrating you're juggling too many
9:02balls you can't be expected to be
9:04specialized in all of them.
9:06>> Yeah. Yeah.
9:07Absolutely true. Yeah.
9:09So you know Aditya so when you
9:12took over the current role it also meant
9:15that you have to you know shift bases
9:18and you know
9:20bring your family back to India from
9:23being in the US.
9:25While we tend to talk about career moves
9:27as if you know they happen in
9:29some sort of a
9:31vacuum.
9:32Uh how difficult was that decision for
9:36you personally?
9:37>> Very difficult. Uh
9:40So I think you know we have moved quite
9:42a bit. I was actually
9:45calculating the other day with my wife.
9:47Uh we've been married 18 years.
9:49And we have moved 11 houses.
9:51>> 11 houses.
9:52>> So we haven't The average is not even 2
9:54years. So we haven't stayed in the place
9:55for more than 2 years right? Uh of
9:58course all credit to her sort of for
10:00keeping up with this and sort of helping
10:02me navigate
10:04uh my career.
10:05Uh
10:07I would say initial few stints when we
10:10were hopping countries and continents,
10:12were quite enjoyable.
10:14Uh
10:15but
10:16this specific move from US to India was
10:20painful to begin with. I had a lot of
10:22guilt, and I think every working
10:24professional would have some guilt
10:25because while from a career point of
10:27view
10:28uh it
10:30it was really looking like a great move.
10:33Uh personally, it was more painful
10:35because kids were of a certain age where
10:37they had stayed
10:39maximum in the US for their lifetime of
10:42all the countries.
10:44Uh and I was worried about their
10:46education, how fast would they settle
10:48down in India, uh what are their chances
10:50of going back and study, and you know,
10:52some of these thoughts you still wrestle
10:54with. Uh but that's that's really uh
10:57part of the journey. Uh but I can only
10:59be grateful to my family uh because, you
11:02know, even the
11:03the teenager son kind of, you know, gave
11:06in, and I think that's what he is happy
11:08now. Uh but it it comes, and I think it
11:10increasingly gets difficult. Uh and
11:12every such move then you have to
11:15watch out for the trade-off between
11:16what's great professionally, but am I
11:18causing a lot of anguish?
11:21Uh and uh is it really good move, you
11:24know, for kids specifically?
11:26>> Yeah.
11:27It's not you alone. It's like the entire
11:30the entire family and everyone. Yeah, I
11:32can understand. So, you know,
11:34uh the most GCC leaders, you know, as I
11:36was
11:37saying, inherit an organization, you
11:40know, and what you essentially inherited
11:42was sort of a
11:44a blank sheet of paper.
11:46And when you arrived, and there were
11:48roughly 100 people at that point in
11:50time, as I mentioned. So, what was the
11:52first thing
11:55you knew that you have to absolutely get
11:58right from the very beginning?
12:01>> So, there were a couple of things.
12:02Actually, I just wanted to uh
12:04extend your introduction. So, we grew
12:07from 100 people to now 2,400 people in
12:10less than 21 months. Uh and according to
12:13me, uh the secret sauce was actually I
12:17would put two elements to it.
12:19One was right from the beginning when we
12:21were brainstorming with alongside BMW
12:24and Tata Technologies.
12:26We placed a great deal of emphasis on
12:29culture.
12:30Right?
12:31And initially when we we were
12:33recruiting, the bar was really high
12:35because we wanted the absolute best.
12:39The bar is still high, but of course now
12:40we have a playbook so the process
12:42doesn't look as intense. We haven't
12:44diluted the process, but we have many
12:46people supporting it now. It looked much
12:48more intense. So we would have four,
12:50five rounds of interviews.
12:52So it would be a technical interview, it
12:53could be a
12:55coding test, coding assignment.
12:58Now imagine that somebody has aced all
13:01of them.
13:02But the final round was cultural
13:04interview,
13:05right? And we were looking for certain
13:07values, you know, is he a team player or
13:10is he a solo player?
13:12Will he be
13:15collaborative? Will he be Will he or she
13:17be toxic? So we looked for certain red
13:21flags because we really wanted the team
13:22to flourish and not individuals you know
13:25taking glory.
13:27And you wouldn't believe Sourav we were
13:29sometimes looking for real specialists,
13:31you know, infotainment security
13:33specialists and sometimes
13:35despite the funnel not being as strong
13:38we would reject candidates because
13:40culturally we thought they wouldn't fit
13:42in. Right? And and that really helped us
13:45because
13:47we believe that you know as a company we
13:50have attracted certain kind of
13:51individuals and selected them and that's
13:54helping us maintain that culture. It's
13:56It's It's not even a soft aspect for me
13:57because you can feel culture day in and
13:59day out.
14:00Another related part was we gave people
14:03a lot of autonomy.
14:05So you can imagine if you're adding 100,
14:06110 people every month starting from 100
14:10or you know, zero to begin with because
14:12we had to recruit those 100 as well, the
14:14first 100.
14:15Uh
14:16we ensure that we get the right leaders.
14:19Uh they get the right team members and
14:20then, you know, like I said, get the
14:22right team and let them do their job,
14:24right? And that's how we scale so fast.
14:27And scaling was one part. You also have
14:29to deliver excellence uh to BMW, who is
14:32our customer and the most important
14:34stakeholder for us.
14:35Uh doing all of this was maddening. It
14:37was very intense. Um as I said, we have
14:40a playbook now. Uh but getting the right
14:43people through the cultural lens and
14:45then give those people high-caliber,
14:48high highly collaborative people enough
14:51autonomy to then
14:53uh
14:54sort of follow through on a strategic
14:55actions and hiring was one of them. And
14:57that's how we sort of scale.
14:59>> Mhm. Okay. Okay.
15:01So, talking about culture at the at the
15:03year, so BMW
15:06uh Tech Works has two different sort of,
15:09you know, very parents you know.
15:13You know, BMW brings German automotive
15:15engineering, premium global brand and,
15:18you know, and Tata Technologies brings
15:20Indian
15:21engineering depth, execution at scale,
15:24those sort of capabilities. How do you
15:25build you know, you just told about
15:27playbook culture. So, how did how do you
15:29build that one culture when you know,
15:32you essentially have two sort of DNAs
15:34versus
15:35>> So, the
15:37objective was to have our own DNA.
15:40Okay?
15:41Now, uh and all credit to our uh board
15:44members. Uh
15:47BMW is very clear.
15:49They don't want a replica of BMW.
15:51>> Okay.
15:52>> Right? And it was very deliberate.
15:55See, any organization as big as BMW have
15:57their own departments, processes um that
16:00serves them well because they are so
16:02huge. Um and so, they need a certain
16:05layers of processes. They need a certain
16:08level of hierarchy.
16:10Um
16:11and that's that's how a truly big
16:14organization would function. They were
16:16very clear that the architecture that
16:18they would expect from us is more
16:20flexible. So, for instance, if we are
16:22running an AIML initiative, it can run
16:24across departments, right? Which are
16:27which are
16:28discrete departments in BMW, but they
16:30would expect that team to work for all
16:32possible departments in Munich. So,
16:34mirroring them was not a great idea. We
16:36wanted to create an entity that is
16:39flexible, that is collaborative, and
16:41that serves BMW's own interest, right?
16:43So, the idea was always to have our own
16:44DNA. We had two great benchmarks in our
16:47parents, right? But every policy
16:51uh role definition
16:53organization structure layers
16:57uh
16:59the way we sort of define
17:01responsibilities, right? Everything was
17:03something that we did on our own. Uh you
17:06know, one thing that we constantly
17:08repeat, and that is something that our
17:09chairman also says, is copy with pride.
17:11So, if there's something great,
17:13we will copy, you know, whether it's
17:14from Tata Technologies, Tata Group, uh
17:17BMW, but it has to be contextual to us,
17:20right? Because we want to be nimble,
17:21collaborative, and serve our
17:22stakeholders best. And it copying their
17:26structure, policies, uh everybody agreed
17:28was a bad idea. So, we wanted to have
17:30our own DNA, and to to their credit um
17:35BMW gave us a lot of freedom.
17:37So, we are not at all a mirror image of
17:39how they're structured in in uh in
17:41Munich, quite the opposite, right? And
17:44uh we
17:46you know, for instance, there are GCCs
17:48where there's a lot of parental
17:51uh overhang.
17:52Uh we don't have that. We have been
17:54given complete autonomy to run. Uh of
17:56course, we align on the basic
17:58principles, deliverables, what needs to
18:00be done, but beyond that, we are free to
18:01run the way we want, so long as the job
18:04is getting done in the way it it has to
18:05be.
18:06>> I think you've already spoken publicly,
18:08uh you know, about
18:10work originating in India, and
18:13your team's contributing to, you know,
18:15BMW's global agenda. So, at what point
18:19does, you know,
18:20uh
18:21an India center, you know, being
18:24somewhere, you know, his work is
18:26allocated uh and become somewhat
18:28intellectually,
18:30you know, from there it becomes a place
18:32where
18:33uh intellectual, you know, ideas are
18:35originating. Is that change happened
18:38already? Where are you on the in the
18:39journey?
18:40>> Yeah, so I think it's a journey.
18:42Uh I think the pinnacle for us, the
18:44North Star is product ownership. Right?
18:47Now,
18:49it's a journey because sometimes we are
18:51transitioning in new topics.
18:54And when you're taking a new topics, of
18:55course, you need some hand-holding.
18:57Right?
18:58Then you reach a stage where you have
19:00autonomy on your deliverables, even if
19:02it's, you know, a piece of software.
19:04Then you may have autonomy on group of
19:06features that you want to enhance.
19:08And then you have the product ownership.
19:10Now, product ownership could be uh a
19:12certain product of an infotainment
19:14system, right? When you have product
19:16ownership,
19:17you're not waiting for instructions.
19:20You're not waiting for requirements.
19:22You define the requirements, so you have
19:23shifted left, you're upstream now.
19:25Because you own the product, you define
19:27the requirements, and then you build it,
19:29release it, right? So, that's the
19:31pinnacle. And since we are new, we do
19:33have product ownership uh in in few
19:35instances, but that's a journey we
19:37track. And that is what makes at least
19:40us unique that there's a defined vision
19:42for us to reach product ownership,
19:44right?
19:45And once you own the product, you
19:47automatically have skin in the game to
19:49be more innovative, to do things
19:51differently, uh to create lighthouse
19:53projects. But that's not the only path,
19:55because, you know, we do have, you know,
19:58few patents filed uh in the company
20:00already. Uh I think um the sense of, um
20:04you know, being enterprise uh is is
20:06already there, because it's also related
20:08to the culture and the lens that we use
20:09to hire people. Uh so, we do have the
20:12autonomy to create lighthouse projects,
20:14um you know, and
20:17and create something that nobody nobody
20:20saw coming, you know, whether it's in a
20:21product or in a feature and then that
20:24becomes a source for global adoption.
20:26So, I I can't claim that we have already
20:29created lighthouse projects, but that's
20:30the endeavor. From a product ownership
20:32perspective, we have few instances where
20:34we have ownership, but it's a journey.
20:36So, I think once you're aiming for
20:38ownership,
20:39everything around innovation gets baked
20:41into it itself.
20:43>> Okay. So, talking about, you know,
20:45patents that
20:47you just talked about, I don't know if
20:49this figure is correct, but already five
20:52patents in last 20 months or so, I
20:54believe. So, you know, that's a very
20:56different kind of a
20:58proposition
21:00from simply delivering projects for a
21:04you know, for just a engineering center
21:06maybe which
21:08you would the way we usually talk about
21:10it. So, how do you create an environment
21:13really for the engineers in India to
21:16actually think about, you know, IPs and
21:19inventions rather than just, you know,
21:21doing the mundane daily work or or
21:24delivering what have been asked for?
21:25>> Right.
21:26>> So, yes, we applied for five patents and
21:29specifically in digital car. So, and in
21:32the space of AI, ML, and automated
21:35driving. As much as I would want to, I
21:37can't take credit for that because
21:40because I think it's really, like you
21:42said, getting the right people,
21:44creating that environment, and these
21:47were a group of six, seven
21:50bright young engineers who thought of
21:53the gap, of the white space, of the
21:55novelty.
21:56Of course, we have a very
21:59defined process in the BMW Group
22:01on how you go about
22:03applying for patents. Of course, we had
22:05made them familiar, but I think it was
22:07really they being really proactive. So,
22:10I mean, if I were to take credit as a as
22:12a company, we would just credit it on,
22:14you know, giving engineers the free
22:16space
22:17and
22:18the
22:20autonomy to choose what they want to do
22:22and sort of apply for patents.
22:25Of course, uh in parallel, we are also
22:28uh creating innovation framework,
22:30innovation policies, uh incentives uh so
22:33that more people get inspired by this
22:36group. Uh we are also giving them a
22:38place of prominence, so they will be
22:39speaking to all employees uh and, you
22:42know, share their experience of how did
22:43they hit an idea, how did they uh go
22:46about applying for a patent, what's the
22:48process. Uh so, we'll of course uh keep
22:50leveraging this, uh but I think the
22:53short answer to this would be really
22:55giving them the freedom and autonomy um
22:57and uh
22:58when we hire, we do look for these
23:00sparks, uh and I think hopefully that
23:02work, but I can't claim credit of
23:04creating a very fantastic ecosystem and
23:06structure that led to this. It was
23:08credit credit to the team. I think all
23:09we could do was give them an
23:10environment.
23:11>> Well, that's very important because that
23:13gives them
23:14the freedom to experiment and, you know,
23:17if you leave a smart person to
23:19experiment, I'm sure he's going to do
23:21something uh good with the time that
23:23he's using for that experiment, I'm
23:26sure. So, you know, for
23:29for more than, you know, centuries,
23:31you know,
23:32you know, differentiation in automobiles
23:34usually was, you know, all about
23:36mechanical engineering, how fast it
23:38goes, what, you know, you know, sturdy
23:41it is and all those things, you know.
23:44But if if software increasingly
23:47kind of determines what the customer
23:50experience would be, then does that
23:52fundamentally uh
23:55change the, you know, automotive uh
23:57industry altogether or a company for for
24:00that matter?
24:01>> Yes, as I as I keep saying, so
24:04mechanical engineering really is the
24:05foundation. Right?
24:07Um many of the brands are built on
24:10promise of mechanical engineering,
24:11right? Whether it's your speed, you
24:14know,
24:14>> Safety.
24:15>> safety um
24:17joy of driving. So, BMW is joy of
24:19driving. So, these are promises made,
24:21you know, on the on the mechanical side
24:23of it.
24:24So, mechanical foundation is absolutely
24:25necessary, but no more sufficient,
24:28right? Because
24:29they say, for instance, that, you know,
24:31the styling, the design, the raw power,
24:34safety can get a passenger in when they
24:37want to buy,
24:38but it won't really hold them anymore.
24:40You know, you need
24:41>> Retention will be
24:42>> Retention will be difficult. So, I think
24:43it's really the software layer that
24:46differentiates
24:47uh
24:48one
24:49from the other.
24:51Uh and, you know, software is becoming
24:53so critical that, you know, people want
24:56the car to be an extension of
24:57themselves, right?
24:59Uh customization, personalization, uh
25:02uh safety, automated driving, people
25:04want every feature possible. Uh one of
25:07the industry leaders, I can't remember
25:09who, but he said that, you know, for
25:10decades we have loved our car.
25:13It's time for cars to love you back,
25:16right? So, which can only happen with
25:18intelligence and and software. And I
25:20think one of the biggest promises why
25:22software is such a game changer is that
25:24it keeps your car current, right? It
25:26doesn't hook you to model year. So, if I
25:28buy a car in 2026
25:31and in 2030, I don't want to look like,
25:34you know, I'm already driving an
25:35outdated model, which was the case in
25:37the mechanical era, right? I mean, I
25:39think software makes it possible for you
25:41to for to be forward and backward
25:42compatible
25:44through features, OTA updation, etc.,
25:46right? So, software increasingly is the
25:48layer where a lot of focus is natural
25:51because that's where the point of
25:52distinction is.
25:54It's mechanical will remain the
25:56foundation, you know, but you just have
25:57to build on top of it.
25:59>> So, you know, talking about this,
26:01and I know that there are a lot of
26:02people who love to have so many features
26:05and being very comfortable while driving
26:07etc. So, while developing these
26:10features, you So, taking that cue and
26:12you've talked about
26:15that the AI opportunity in engineering
26:18goes well beyond, you know, just
26:20writing, you know,
26:22codes faster. So, where do you see the
26:24much bigger productivity,
26:27you know, breakthrough across the entire
26:30engineering
26:32life cycle if I may put it that way.
26:34>> Yeah, I think uh my comment was really
26:36when we look at AI productivity
26:39the natural focus and disproportionate
26:41focus is on software development and
26:43coders, right? But writing
26:47millions of lines of code extra is only
26:49going to create bottleneck down the line
26:51and we don't even have capacity being
26:53the sense the industry may not even have
26:54capacity to sort of validate those
26:56codes, right? So, the the promise of AI
27:00can be leveraged when we look at it from
27:03end to end. So, across the software
27:06development uh life cycle, yeah.
27:08And I think everybody is sort of trying
27:10to feel their way.
27:12Uh so, right from defining the intent
27:15and the requirement architecting it
27:18coding it, building it, testing it,
27:21releasing it, right? Now
27:25the
27:27challenging bit
27:28is
27:30in industries like automotive
27:32is that we deal with a lot of safety
27:33critical applications. Right?
27:36So, no matter how much you rely on AI
27:39uh you will have to do physical testing.
27:42Right? Because uh if you
27:44uh
27:45develop a certain code it has a bug in a
27:47certain application on your phone it's
27:49abstract, right? But uh in automotive
27:52world it could be fatal, right?
27:53Depending on the application. So, it
27:55comes with a lot of testing whether it
27:57is hill testing, whether it is on fleet
28:00testing, uh you know, uh
28:02so it's it's really something that you
28:04can't avoid in automotive. So,
28:07I'm not giving you a straight answer
28:08because it's still work in progress
28:10according to me.
28:11>> Let's put it this way, which part of
28:13engineering
28:14auto engineering, automotive engineering
28:16do you think that AI will change things
28:19dramatically which we are not expecting?
28:21Probably some of those are
28:23which it will not as you said, but there
28:25are some things which can't change. Do
28:26you think
28:27>> think yeah, see for instance if I look
28:29at software development life cycle in
28:31general, you know, the way we want to
28:33measure it is, you know, from
28:35requirement to resolution.
28:38You know, this whole cycle time will get
28:39crunched.
28:41The only point I'm trying to make is it
28:42could be contextual. So, for instance,
28:44if you're releasing features in
28:45infotainment, it may be much faster
28:46because you can live with bugs if they
28:49pass into the production system. Whereas
28:51for something like A dash, you wouldn't
28:53take that risk, right? So, it's
28:54contextual to the field and domain
28:55you're talking about.
28:57>> Yeah. No, so I I can understand that
28:59because
29:00you raised concern about AI agents
29:03kind of, you know, acting in ways their
29:05creators didn't explicitly
29:08ask them to to. So, in a car where
29:10software decision can, you know, have
29:12physical consequences, do you think
29:14there is a point beyond which you would
29:17not like AI
29:19to have autonomy?
29:21>> Absolutely. I mean, at this stage I
29:23can't imagine why would I trust AI
29:27without, as I said, doing the physical
29:28test. In fact, it's not just you know,
29:31it's not just whether OEM wants to do it
29:33or not, it's also regulation, right? So,
29:35the standards that are defined,
29:38sometimes you want to trace it back
29:40to the logs and, you know, find out
29:42where the bug was.
29:44You know, some of the hill machines are
29:46actually preserved for a number of years
29:48because in case of an accident, you want
29:49to trace back to the
29:50exact test case, right? So, if the
29:52stakes are so high,
29:54I don't think AI
29:57in the foreseeable future will actually
29:59replace the validation standards that
30:02auto OEMs have.
30:03So, we'll have to take it with a pinch
30:05of salt, the whole promise around AI.
30:07That's That's my personal belief that
30:09it's it's too critical
30:11and I don't think we have reached that
30:13level of accuracy, confidence level for
30:16us to dial down
30:18the physical testing, especially in
30:20safety-critical application.
30:22>> No, true. I mean,
30:23here a human life is at stake and,
30:26you know, the other sector where I see
30:28AI being, while of being very of use,
30:31but still people are cautious is
30:33banking, where only money, human money
30:35is it's not the life after all, but then
30:38here we are talking about life, so yeah,
30:40of course. I think the final question,
30:42you know, you
30:44you know, when you're building something
30:46this quickly as you said in
30:482400 in 21 months, so my numbers were a
30:51little outdated, so everybody comes to
30:54the CEO for, you know,
30:56certainty, stability. So, but there must
30:59be moments when you don't know the
31:00answer either. And you said that your
31:03job is to ask good questions, people
31:04other could be, you know, as good as you
31:06to ask good questions.
31:10Who does the CEO go to
31:13when, you know, when when he's he's
31:15unsure?
31:16>> Yeah, I think as they say, it's kind of
31:17lonely at the top. But honestly
31:21I didn't feel that alone because and
31:25it's counterintuitive, but some of
31:29all of my board members are great
31:31mentors.
31:32Right, so
31:34and they have different specializations,
31:35right? So, if it's a topic related to
31:40if it's a
31:41topic related to people for instance, I
31:42know who to go to. If it's
31:45topic related to product strategy, you
31:46know, I can bounce off ideas, so you
31:48know, their job is to challenge me. But
31:51I'm I'm lucky to have them also as
31:52mentors because I can also discuss
31:54problems with them. So, I think
31:57the board is extremely supportive, very
31:59ambitious,
32:00aggressive, but very understanding in
32:03terms of discussing constraint and sort
32:06of, you know, basically
32:08nudging me the right way.
32:10Beyond that of course uh
32:12you know, my family.
32:14So, my wife,
32:16of course I can't discuss everything
32:17confidential, but she knows when, you
32:21know, a cloud is sort of hanging over
32:23and sort of, you know, she tries to
32:24troubleshoot with me. And she of course
32:28the best thing is she is not colored or
32:30biased because she's not in the midst of
32:34it, so she can give a very neutral
32:35feedback.
32:36She can be an outsider's view.
32:38Outsider's
32:40very dispassionate right?
32:43Brutal and I'm like
32:45for instance I'm playing see this is the
32:47situation I'm dealing with and I can't
32:49understand why
32:51the other person wouldn't get my point
32:54of view and she's like of course because
32:56you're wrong in this situation your
32:58point of view if had I been there this
33:01is how I would have thought so why don't
33:03you change your
33:04posture essentially. So of course she
33:08helps I'm very close to my sister my mom
33:11so you know bouncing off ideas of course
33:13you have to be very careful because you
33:14just give them a hypothetical situation
33:16to react to because you don't want to
33:19burden them with details and it's not
33:21even proper.
33:23But yeah and I think I find a lot of
33:25inspiration in books in in in
33:28documentaries in sports so
33:33>> All right Aditya thank you so much for
33:35speaking to us and I'm so glad that
33:37stars have you know aligned for you
33:40yeah we could finally catch up thank you
33:41so much for for doing this thank you.

About this transcript

This page contains the full transcript of The Art of Building a 2,400-Person Organisation In 21 Months Ft. Aditya Khera, BMW TechWorks India by ETGCCWorld , generated from the public captions YouTube serves with the video. The transcript has 5,412 words across 935 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.