Subscribe
Share
Share
Embed
‘Innovators Playground’ is a podcast series where we explore the cutting-edge innovations and insights shaping the future of technology. It brings conversations with leading founders, industry experts, and thinkers to inspire and inform youth around the world.
DILIP: Welcome to Innovators Playground, NPCI's tech podcast created to inspire engineers to come forward and build for India. In this podcast, we speak about Breakthrough Innovations, DPI's that India has built, some of the great platforms have been built on top of DPI's to serve millions and billions of Indians to move forward to the digital financial services. Let me introduce Lalit Keshre, founder and CEO of Groww, who is an IT engineer and a hardcore techy, who has worked tirelessly for the last 10 years to build the group. Lalit, welcome to NPCI's podcast. Thank you for agreeing to do Innovators Podcast for NPCI. Let's start the session. How do you see the open source value, open source contribution to Groww? And how do you actually look at this whole open source revolution, evolution, if I may say?
LALIT: First of all, thank you so much for inviting me on the podcast. On the open source thing, we use, of course, we use a lot of open source, of course, the enterprise version because you need kind of, you also in one way support open source web, taking the enterprise versions. But I don't think Groww would have, we would have been able to build the platform without open source, like right from frameworks like Kubernetes or Kafka. And since the beginning, our kind of platform has been built with that. The way I see it is that the tech innovations mostly happen by developers and people who want to kind of contribute to the world and so on.
DILIP: When you look at, you know, open source brings in its own issues, right? In terms of the community and suddenly we see the risk of somebody, some big giant buying them over, right? And again, you suddenly from open source, you're forced to the license version. So how do you strategise? How do you select? How do you look at those issues?
LALIT: So this is a very good question. So, I think we always chose frameworks or any tool which had a very solid historical kind of, so you can't test out any new stuff. So, in the tech world also, there's like every day you will see some new thing. Coming up and so on. So, by design, because we are so and for us like reliability, security, everything is just so important. We can't just use any untested stuff and so on. So we went with (different) kinds of things, for example, Java. So, we started with Java, right? And it's like there's such a huge community of that language. All the tools if you talk about right, like we use all well-tested open source tools. We made some mistakes in early years, but later on, I think we kind of moved back to everything very very (quickly).
DILIP: How's your personal interaction to open source? How did you get there?
LALIT: To be very candid, like I and now I don't code anymore. I mean, of course, I do vibe coding on the weekends, you know, write apps and so on. So, I started as an engineer. I coded for like maybe around eight years or eight to ten years first in hardware languages. So, I was a microelectronics engineer. So, we used to code in these languages called Verilog and VHDL. You write chip design and so on. And then when I did my first startup, I coded in Python and used… again, there also we used open source frameworks like Django. I don't know if I mean they're very popular at that. It's very easy like anybody can get started and so on. I've never been so into contributing to open source, but I remember one incident that when I was studying, we used to design these PCBs, right? Like I was in microelectronics and when you design a PCB, there is a file that you generate after designing which people can use again. So I don't know what happened. We used to work in the night like, you know, late nights and we used to call (them) night out. So, all the night you work and if it worked then we uploaded that image on a website. The next day we saw like there were thousands... now we talk about millions, but there were like thousands of downloads and I was so elated, so happy. I did not get that happiness even while competing that project, but I got more happiness when people downloaded that image which was like…
DILIP: See, I started Open Source contribution journey very early.
LALIT: Yeah, but I think I need to do more as a company also, we need to kind of contribute a lot more. But the growth, I mean touchwood growth has been very good and so on. So, we've always been kind of struggling to manage the scale and I don't think we got any breathing space. Okay, we can think about it. Let's now start contributing to open source. I hope that those times will also come.
DILIP: You know, at NPCI also we are trying to incentivise people basis now as a key criteria to the open source contribution now. So those are the kind of changes at NPCI we are also doing, you know, and I think this moves us to the next question, right. When you started Groww, you know, what were the first principles for the tech choices and how do you when now you reflect back how you have fared? How did the right things happen? You found the early mistakes you made, a kind of chain management and revisited those assumptions and or principles.
LALIT: Well, that's a great question. So, I think I'll tell you maybe some good decisions we took and then some mistakes that we made. So, one good decision was using very very… as I said like very very time tested, proven frameworks to start, to build everything right. So, by platformisation, having a distributed architecture, our org structure is also like a very distributed pod structure and kind of that also maps well to our tech structure so that people have independence and they can move fast and so on that also created some problems later on. I'll come to that and then also another thing was investing heavily in our in-house staff. See nowadays, you can build a startup by kind of okay getting this third party service, this third party service and you can assemble this and kind of launch very fast and so on. We realised that the scale that we were imagining I think would fail like in all those B2B third party services. They typically are good if you are running like a small kind you know at a small scale and so on. We wanted to scale significantly. So, we invested from day one like we invested heavily in our own infrastructure in our own kind of services and so on. So microservices. So, those turned out to be good. Now what did not work out is and that's also like around 2022 we realised around year 22 that we kept on building so it is all distributed but it became like a distributed monolith. What I mean by that is like there is a microservice running here. There are others, but then there is just so much interdependence. It's like noodles like you know everything and so on. So, although it is distributed it is still kind of like a monolith and the scale, like growth was very fast and we are very grateful like we, I mean nothing like growth. Growth kind of teaches a lot more but then we kind of pulled our socks and like you know kind of tried to kind of simplify it and move towards simpler platformisation and so on. Another mistake was the platform choice for the front end. You know, if you want to launch fast, you build up, you pick a framework, ke wo Android pe bhi chal jaaye, iOS pe bhi (that it should run on Android as well as iOS), and then web and everything and then it's suboptimal because you don't utilise the power of the platform. It was another one year cycle to get to those native components and all those kinds of things. So, overall I also feel that you never get done. So, because you make some decisions and those decisions last you for maybe three or four years and then there is no point in building for 10 years down the line. You build for next like two or three years and then again you keep building. So, I have an opinion that tech debt is always good. There is no point in building everything and then there is no customer.
DILIP: I think it's a great point and you know we at NPCI also process large scale into about 15 billion APIs daily and you know I think the whole conversation is about you know how are you ready for the 10X right because this whole open source story works very well. You know I'm saying look at the in-memory or Kafka message you are or a high speed DB but when you get them together right on a completely commoditised hardware right it has to function at that scale. You know in a horizontally scalable manner. I think this you spoke about this simplicity right. And you know maybe we could talk about some of the things on how you ensure that simplicity is protected when the new code is written. How is your whole DevOps site with the code review or the DevOps cycle or the change management cycle is tied to your principles, what you spoke about.
LALIT: Simplicity is the most complex thing I feel.
DILIP: It's easy to speak about it.
LALIT: Yeah it's complexity is like gravity like no matter how simple you start you will keep getting more and more towards complexity. What we did was again credit to Neeraj and his team continuously fighting this complexity thing and so on. It starts from the product side right. So, now the thing is that you can build everything. One thing that we realised is that you don't need to build everything you need to build what customers really want, and that kind of solves some of the problems. Second thing is you build an architecture which enables people to build whatever they want. The way we have designed the org is like we have pod structure. Now there are more than 10 products we have and they are very very small teams like startups running on those products but they don't build the infrastructure. So, there is a platform layer which is common across everybody. The job of the platform layer is to kind of manage scale, provide what basic services these application folks need and you know let them ship faster quality code and so on. But then they run independently. So that is how Neeraj and his team build the architecture.
DILIP: So when you're seeing the platform layer so whatever resource you are using there is a layer which has been built by you. So that so for example there is a layer for Kafka. Yes there's a layer for databases.
LALIT: Everything, everything, so database also example we build ourselves. We don't use any again third party kind of invested a lot of time like I think an year or so to build that ourselves and that…
DILIP: Fantastic, that would bring some sort of a standardisation and while you are building the platform layer it also takes care of the resiliency part or the error management or this kind of things which is which otherwise every team is solving. That you have kind of ensured that; so, what you're saying is let me again try to understand. So, what you're saying is, most of the developers are building for the customer, what customer wants.
LALIT: And then there is a very lean team like there is a lean team which is in charge of the platform. And in platform there are some use cases which are like a lot more application side also which are common across all the pods.
DILIP: So, every engineer is actually contributing back, adding more experience to the consumer or a feature or…
LALIT: So they're like customers to this platform team. They are for the platform team. We need this much scale. We need this much capacity and those kinds of things.
DILIP: Yeah fantastic. I keep talking to this too many tech founders, CTOs who were dealing with scale. How do you look if there are problems right and when you build such a complex infra the observability becomes a key issue. Right. I'm saying exactly where exactly the problem is. How do you deal with that aspect because finally it just comes back to a single platform. Right. I'm guessing while multiple ports are coming somewhere it converges back to the single app. Yeah single platform and then you talk to the market. Yeah, whether you talk to the UPI system, whether you talk to the NSC BSC the trading platforms or the mutual fund houses your internal platform and the external ecosystem. Right. Both have to work in complete sync to deliver that experience.
LALIT: Yeah. I think very so I have two points here, like one is it's an observability is something which you don't think about when you kind of you know when you start.
DILIP: That's a mistake.
LALIT: That's the biggest mistake that you make, and then it hits you like you know, let's say you're one million traffic coming on the app even if you have designed observability for let's say one million traffic at 10 million you will fail so you have to continue investing in observability and these realisations come little late like you know, you're a startup you start with 100 users. You say okay 100 to me. I am checking like when we had 100 users I would check like directly, you know on the DB and you will write a query and you will and you don't have anything much to do. So, you keep writing queries and seeing various kinds of patterns and so on then your thousand then you have 10,000 and so on. The second thing is this third party thing right. So, our systems are very complicated in the way that we depend on a lot of third parties. If you don't have the right observability system, I think you will spend like 70-80 percent of time just finding out where the problem is. We had our own share of challenges but as they say like you know, you learn as as you go.
DILIP: Brilliant. You know, some of the thoughts that you're reflecting on exactly match for NPCI as well. And I think it's fresh coding required. Your segment is fairly diverse right now. You are building for the complete demographics the complete even from the income standpoint.
LALIT: Yeah, right.
DILIP: You're truly making India kind of a company, you know, supporting 50-100 million consumers with such diverse needs and demands from the platform. How do you design something for such a large number of stakeholders who has very different ask.
LALIT: So, that's the kind of abstraction that we kind of try to do. So, let's say other user layers like there are some use cases like one is like as you said, like there are a lot of HNIs on the platform and then there are people doing like 5000 rupees SIPs or 1000 rupees SIPs and so on.
DILIP: Or now 100.
LALIT: 100 it's still like we are not there. I mean, yeah, we are still not there yet. I would like I think once we get to 100 rupees as IP will have like 200 million or 300 million MFs and investors in the country.
DILIP: Potential 10x story.
LALIT: Yeah, so that's the thing then there is an activity level. Now there are people let's say I'm investing you can do it in few milliseconds, but then there are like high fequency like traders there are like millions of traders at the same time buying some you know, either they are trading derivatives or intraday or whatever and it has to be microsecond kind of stuff. Right. And then there are other constraints like they're like multiple features. Some features are very high DAU. Some features are like low DAU. Yeah, some features need to have very high reliability like payments for example, or this order management system or risk management system and they are some like cosmetic. Okay, if you like, you know, if this logo does not look good, the customer does not lose money, right and so on. So we have this framework where you know, what is the order of magnitude in terms of customers level and then there is this which is the intensity level which is like the field of pain and then reliability matters there. Another framework that we have is 0 to 1, 1 to 10, 10 to 100. If you are doing any changes in 10 to 100 like it goes to a crazy amount of review process, right? It has to kind of everybody is involved there a bunch of reviews 1 to 10 again, it is basically kind of it can impact like millions of customers. If it is 0 to 1 you are testing it and it is not involving customers' money in those kinds of user flows, then you can move ships fast. So, shipping velocity is different across this foundational team. We talked about how the infrastructure moves very slowly. Yeah, they need to move very slowly because they can't keep deploying every day. But the top layer guys can keep deploying every day because there's a small use case you quickly test and deploy.
DILIP: Fantastic. I think this I love the idea of you managing this velocity of the people, how fast they can contribute or how slow they should contribute to the platform.
LALIT: Another thing is that there are people meant for that kind of thing. So, if I look at my behavior, I can't be in that like, you know, infra team because I am hyper. I need to ship everything every day. Right. And so on. And those kinds of folks are they think scale, they can foresee scale. And then there are these folks who are like customers. Now, I want to ship it tomorrow and those kinds of so they are very different attitudes of people also.
DILIP: This is my favourite thing. You know, I keep talking about there is so much for India to build right now. And when I look at our engineers, you know, we have built India's stack. You know, Nandan has been involved. He's been a kind of guiding force to build this whole Aadhaar stack. Then he's with us to build a payments stack and a variety of new things are happening. How do you see that, you know, the next level of DPI is on how do you foresee that? What's your advice for the engineers that what do you think is the future in this whole DPI story or what are the still pain points in the ecosystem which needs to be solved either by the startup or by fintechs or by the entities like NPCI, UIDAI or ONDC for that matter.
LALIT: Since we talked about DPI, we owe a lot to DPI. Groww would not have existed without it, so be it, you know, digital onboarding, Aadhaar, KYC, UPI. We were the first ones, by the way, who supported UPI.
DILIP: I was actually encouraging you to get there.
LALIT: Yeah. And so, DPI is a platform where so again, since we are talking about platforms, DPI is like a platform which enables startups like us and hopefully it will enable like, you know, hundreds of more startups that will grow very, very large and so on. The second thing is what challenges we see, Nandan envisioned this, Dilip you’ve built NPCI and so on. I think we have got a good start with your leadership. If we can have, you know, 10 more Dilip’s and Nandan’s, then I think we will be kind of, you know, like just imagine today UPI is like, you know, is a role model for the entire world, not just for not just for India.
DILIP: When we were talking to Nandan and team that, you know, they were doing virtually the entire D-loop in just like 24 or 36 hours.
LALIT: This is marvellous engineering. This is like crazy engineering, right? Even UPI, for example, the number of QPS or the number of transitions. So, what I'm trying to say is that we are so it's not like we are behind or anything. We are there. The talent is there and so on. And then we have visionaries like you folks and so on. I think we are so well set and we can continue building and especially with AI. I think I agree with what Nandan says. There is a lot will be done. This is the application layer or, you know, orchestration layer or anything. So, the developer talent, the number of developers that we have, I think nobody can compete with that right today.
DILIP: You know, I think let's shift to AI. You spoke about AI and at NPCI, before you tell your views, I want to just tell you NPCI has been building models for the last four, five years. You know, every transaction process by NPCI has a risk score. It's a real time inference about 40, 50 models for the fraud risk mule. And somewhere I kept thinking that we were very good on AI. And but last one year we are completely blown away with this whole revolution while Open AI has been doing this for the last three years. But the last one year has been completely mad. We are also reflecting on what we are doing in AI and whether it's an incredible job which will look at NPCI for the AI work while some things are in progress. How do you see this in Groww and maybe you could elaborate your vision. What is the current state, future state of the use of AI or the disruptions that AI is going to bring in?
LALIT: The speed at which things are happening is mind boggling. I'm very paranoid about how it is going to disrupt everything. I think it is going to disrupt us like individuals. It is going to disrupt organisations in a very different way. And I am still kind of trying to figure out but few things are clear. I think the speed now I can say in our world, like, you know, we used to ship like, okay, there is a 15 day, you know, two weeks print cycle or one week cycle. And that is dead like that's like gone. Now people are shipping like every day. We used to talk about, you know, promotion from SD1 going to SD2 to SD3 and then manager and so on. I think that is dead like that is done. I think there will be much, much fewer layers. We used to talk about job descriptions like, okay, there's a product manager. There is a designer. There is analytics. There is engineering. In engineering there's a front-end engineer, back-end engineer, QA quality engineer, there's a front-end engineer.
DILIP: Database expert. You know, I think variety of roles we created.
LALIT: Exactly. So now I think all done like, you know, all done. So, speed, the org structures, the layering and everything, the functional, what is it called like, you know, functional boundaries and then automation like, you know, coding. So, people say it will disrupt engineering and so on. I think it will take some time, but coding is different from engineering, right? One thing what will work for at least for the next few years, we don't know when AGI and all those kinds of things will come, but thinking, creativity, imagination. AI still like today, average of whatever they are kind of training the models on and so on. But if you want something outlier, you will have some job there. I know,
DILIP: You know, while our team still kept on insisting on this deep expertise on a variety of different open source modules and I kept on telling them the only way we will survive is the full stack developer. And now it's a full stack AI developer. Right. I'm just saying that's what I think our requirement is just a single job description. Yeah. Are you a full stack AI developer? Yeah, then come inside and it's going to be easy, right? You know, and we have been chatting about it, you know, the some of the experiences you were narrating that, you know, in the Silicon Valley, how people are building. Maybe you want to talk about that, you know, the youngsters are building. So yeah, the way you spoke about the experience collapsing, the designation collapsing, the function. I love it, actually, the functions collapsing. So how do you see the future or what should this young engineers do while they are doing the engineering?
LALIT: Yeah, I am actually not worried about young folks, young engineers. I'm worried about the old folks…
DILIP: like us.
LALIT: Like me, because I think young folks are faster in adopting things and so on. I think I was talking to a startup founder just two days back, and he said that they are experimenting with hiring high school grads. Which is crazy. All the lot of startups that are being built as startups that are being built in the Valley are like 20 in there in 20s, mostly 20 to 30s and so on. So I think to be kind of, in short, I am not worried about young folks.
DILIP: But I but I think the point, Lalit that I was saying that, you know, if you know, the formative four years or five years of engineering, a lot can be done. While doing the engineering. As a country, if we want to build a lot right in the next 20, 25 years, there's this four year cycle or five year cycle. Right. And now you're talking about from high school itself. Right. So while preparing for your JEE’s and everything, I think this whole story of six to seven years, what people spend including their engineering cycle is going to be very formative. That what they want to build for the country. Right. Yeah. It's so fascinating to see that the country needs so much. There is a huge demand on how much India can do and what India needs.
DILIP: And we have really great engineering talent now in the pipeline. But if what should they do actually to to really ensure that they're one is their aspiration, their skill set and their actual outcome matches to what India needs.
LALIT: So, I have three things. I think one is be curious. I can give you some questions. For example, if I would think about so we talked about Aadhaar architecture. 1.4 billion Aadhaars D-loop itself. I just think about it. How do they do it? Right. You know, so how does NPCI work? How do they manage like, you know, QPS and how do they kind of maintain the reliability? Like, you know, if you are in a shop and paying and then if the act comes like a two minutes delay, like the transition would not happen. Right. So, those kind of things would be curious about how the world around, how the LLM works. So now people are using LLM, but like it's so phenomenal when you read about how it really really works. Right. And so on. So be curious, number two, create something when you do something, then only like something kind of, you know, like you get the feeling. So, like, you know, you can't swim just using to feel the water when you're coding. You need to really feel how things work and so on. Right. So that is second and third is I think don't worry. Like I think I see a lot of young folks coming to AI is coming like AI is coming like, you know, Game of Thrones. People keep saying winter is coming. People are saying, oh, yeah, it's coming like AI is coming. So, like, but if you keep worrying, then you will like you will be kind of bulldozed. You're wasting your energy because you can't do anything about it. It is coming. People are all going to stop working on it. So I think these are the kind of things that if I get back into my college, I would do this. Yeah.
DILIP: Yeah. I think we spend a lot of time doing Pascal programming and all these things, giving exams. Fortran. Yeah. Same for a stream. Likewise. You know, when we spoke about AI, how are you actually using it in business? And does this whole, you know, this in the LLM architectures, the cost of inferencing and you know, the overall, you know, how are you building the fully AI powered Groww?
LALIT: So, we look at in two ways. One is, of course, from the consumer side of things. And second is on the productivity side of things. So, in consumer side of things, the simple question is how can you evolve the customer experience, right? To personalisation or like a bunch of things there. And on the productivity side, of course, how can we be more efficient, faster, higher quality and so on. And these are the two kinds of frameworks. On the cost side, currently, my thought process is that let's not worry about the cost because eventually tech like when I was in college, the Internet bandwidth was just so expensive. And if I were to optimise, let me kind of make it cheap sometime, maybe so we are anyways converging to like, you know, very, very small, marginal kind of cost world and so on. So, if you just compare the token like cost of AI, like one year and now it's like phenomenal difference and it will continue. So don't worry about it too much. That's kind of hurting us a bit in India because not from the cost perspective, but the leverage of AI versus like, you know, innovation, AI innovation. You can always put like, you know, you can always set up like non-automated way of doing things because it will still be cheaper in the US. It's not so. And I think that's one of the reasons why that AI adoption is much faster in the US. I am very optimistic about AI and India.
DILIP: So, tell me what's the culture at Groww, you know, when an engineer is in your office, how's the culture? Do you have games? Do you know how you deal with this whole you spoke about the high productivity, deep tech outcome? Right. You know, how do you create that around Groww?
LALIT: The most important thing is how it starts. So, basically what we have learnt is that the existence of grow, right, the existence of grow is to basically democratise wealth for India. And if you just understand just one thing that it is not possible without tech. So, it's the backbone. So, once you understand that, then everything revolves around that. And then if you look at the top level, also like I've been an engineer, all my three co-founders were engineers, although we wear different hats. I don't do engineering anymore, but I wear a product hat. But like we talk about day in, day out about that. So, tech enables access. So, let's say today, even if you look at today's penetration in India, look at any product investing, credit, insurance, UPI, of course, you have done phenomenal. You are top of the funnel for all of us, but it's not possible without tech. So, when it becomes like the backbone, then you have that's the kind of engineering becomes the becomes the core. And then a lot of rituals like, you know, engineering is not about coding. That's like the biggest misconception. Like what I have learned is that you spend less time in coding nowadays, like coding is anyways getting kind of disrupted and so on.
DILIP: So how does the office look like? I'm just saying that there are games there or…
LALIT: it's yeah, there are games. I mean, it's been long. I went to that part of the office, but there are like a bunch of games people. So, we have worked from home. work from office. We don't do work.
DILIP: same like a NPCI. And I get a lot of beating from NPCI employees. You know, if you call yourself a deep tech company and Bangalore has been a big spoiler on this part, you know.
LALIT: Yeah, but we have been very opinionated on that. Like we were work from home and there was pandemic and so on. But what we realise is it becomes very transactional when you are work from home. It's like I will tell somebody this and then there will be like, you know, it's like more transactional when you're in office, a lot of creativity, a lot of and in the age of AI, I think creativity and insights and you know, those are big thinking, big, those are the only thing that will remain with us work from home. I think kills everything. It works very good for like an MNC and so on.
DILIP: And the BPO and I keep telling my guys also that we are not a BPO and when you are with people, your learning is 10X. Your interaction gives you high energy. You know, you drive energy from people around.
LALIT: and I love the energy like you go in the office and then if you see like empty, like everything, it does not feel like you have to talk to people. I will talk to engineers and so on. Right. So, it's like you need that energy.
DILIP: In fact, in NPCI, we have a policy now, 8 is late. So, actually the guard takes the people out at 8 o'clock because you know at 8 o'clock, if you come to the NPCI office and you will see such a high energy conversations, calls.
LALIT: So, in Bangalore, you can't do that. But we are very flexible. So, people, so although it is work from office, but then people can choose to let's say there's high traffic, there is rains and so on. You can work from home. So it's very flexible and then there are discussions. So ,we have like lot of all the time. I see if you walk on the floor, like people are discussing, talking. I love that kind of energy.
DILIP: I can't let you go because you know, you come from a small town in Madhya Pradesh and how did this whole thought of IIT and engineering occur to you? I don't know how is the surrounding. Yeah.
LALIT: So, I see I've been very lucky since my parents stay in a village, but then I started staying with my grandfather, he influenced me a lot and so on. So, the English medium school just started. So, I was lucky that it was starting at that time. I got into that and then around I was in 11th grade or something. I saw a library and I was generally I used to read a lot in library. I saw this book called IIT JEE, I asked my principal, teachers did not even know. So I asked my principal. He told me that these are the most prestigious, very tough kind of to get into so on and so on. So I kind of started kind of reading and then my grandfather's boss used to get this newspaper Times of India. It was one day late because you get the same day Times in the smaller city. So, he used to nobody used to read there. So, he used to get me at my home and then I opened times and then I saw the ad of Brilliant Tutorials. Chennai based postal course. Yeah, I still remember the address because I wrote them so many times. So I started those that postal course. And I was good in math. I hated the other subjects. Now I did not hate but I did not understand physics chemistry too much. But math like was my strong area. So somehow kind of I got in.
DILIP: It's very inspiring that you could do this at you know when because today the cohort right without a cohort without anybody else.
LALIT: At that time see I think to be fair like now it has become very competitive. I think when we wrote when I wrote JEE. I don't know the numbers now but it was not so competitive like how it is today.
DILIP: Still I'm just saying that working on this postal arrangement and I think it's so inspiring. Yeah.
LALIT: That was a very good course.
DILIP: Again. I think your kind of ambition and supported by the learning you're calling it luck. But I think there is nothing called as luck when…
LALIT: I believe I mean if I look at my life
DILIP: Your ambition, your efforts, and your desire and how you focus actually. How did Groww happen? How did multiple startups you got into? How did you think about this idea and how did you pursue?
LALIT: So Groww started in 2016. I got out in 2004 from IIT and in IIT I got excited about investing in stock. So, you don't know about the stock market or stocks or so on. One of my wing mates got into internship at some company and I got to know it is listed. What does it mean that you can buy share, 80 rupees share it was and I got like super excited. You can own the companies. I used to get some stipend. I got like 80 rupees. I don't know how much I bought but and then it was state. I got got into this rabbit hole and then on the other side, I was doing engineering like I was like I was doing digital signal processing video, video algorithms and so on. So I started with this startup called Eduflex which is like JEE the way I was preparing was very painful,no teachers and so on. So, you create like a education tech company where teachers could create videos and stream and so on. Did not last. Joined Flipkart. So, coming back from when we were at Flipkart, we started thinking about what next. Education I did not do. I did not want to do again. Yeah, I still feel that you can build like still still there is a big scope.
DILIP: It's a hard it's a hard problem to solve.
LALIT: It's a hard area. Everything is a hard problem. But I think the three years of kind of you know that gruelling sucked out energy from me. I thought I'll do something else. And that experience of opening the account so many years back, stayed in your mind. And then I met Harsh, Ishan, crazy set of folks. I mean, again, I'm lucky there that I got in touch with them. Somehow, we started and financial services looked like very ripe. So, this time also I was more logical like earlier. It was more lot more just pure engineering thinking and then experience with Flipkart failure of startup that got me some product and business thinking as well, you know, you look about. Okay, what is the market size? What is your right to win? What how will you play and so on so that's like we thought you know.
DILIP: Actually just it just struck me you're you're part of that Flipkart Mafia gang. I'm saying many startups came out of Flipkart. Yeah, I think it must be an amazing journey, but I think this is whole, you know, the investing, the rules, you know, because starting something new and I'm sure some of the other startups would have felt right. Creating a fintech in a regulated area. Would need a lot of understanding of the space. How did you get that business domain or how could you survive in that?
LALIT: I'll tell you one very simple thing. I think we were stupid and that's why we did it. Like I think because of that we were successful. We had absolutely no idea of the complexity. If I had known Dilip that you know, it is so complicated, I would probably I would have done something else. So, I mean, just being very candid, like we were a bit kind of ignorant in my like and we also we were also like we're enjoying the process like that was the most important for the first four years. By the way, we were like a zero revenue company and you're just kind of doing your coming to office every day, having fun, customer feedback. So one best thing about us was like it worked like, you know, for first one year, we were kind of dabbling into multiple ideas and so on. But when it started working, the customer feedback and customer interaction is the best thing a startup can have and then it gives you direction to kind of keep moving keep and it becomes very exciting. I mean, you keep learning regulations came like later, but before that domain itself is so complicated and then regulations come and then business like, you know, making money. We did not even think that when we started with that. Okay, we'll just like internet model, right? We were like kids of internet models and so on. We thought, okay, we'll figure out something.
DILIP: I think the principle of building engagement first, I think pays off is what I've seen. If you're able to build engagement, I think the revenues and the business model emerge over the period.
LALIT: I am a strong believer in that. So, engagement plus so there are four things in my model like, you know, acquisition. So, you can't be spending money and acquiring customer. Engagement of course, engagement tells you how useful you are to the customer. Third is customer love, right? You know, if customers still don't love you, you can become replaceable and lastly, Retention - like how long can you kind of be with the product? Right. So that's why UPI is a great product. You always you'll always pay somebody.
DILIP: Yeah, no, fantastic. Let's move on to some fun thing. Rapid fire questions. One financial mistake people make.
LALIT: The biggest mistake people make is that they think that it is exciting. Wealth building is exciting. You need to trade. You need to kind of be there. Right. Lot of wealth is built by just being boring. Just watching it. Do SIP. Just keep watching and don't kind of hyperactivity is negative.
DILIP: One of the book or podcast you would like to recommend to the engineers or techies.
LALIT: So, I would recommend two very extreme books. One is on the creativity side, which is a Lord of the Rings, which I read in my college and like it opened like a lot for me. Right. Crazy, crazy. And I recommended this book to my daughter and she took like three years. So, kids don't listen to their parents. So, it took her three years, four years to kind of listen. But once she got hooked into it, she's like no fan and so on. Right. So that is one and second extreme side on the engineering side is my co-founder Neeraj recommended this like just around six months back, which was AI engineering. It is a beautiful book. Very well-written. I don't typically read tech books, but this one I read and found it very, very good.
DILIP: Fantastic. Your favourite movie?
LALIT: Lot of them. Like I think I am. Nowadays, I don't get to watch, but I just recently I watched a movie called Network. It's 1976. It won a bunch of Oscars. Yeah. It kind of tells a today's story like it's so far. It's so much foresight in that movie. Phenomenal movie.
DILIP: The next big tech innovation you think?
LALIT: I think understanding human behavior. Right. I think in a productive way. So, some folks have understood and created reels and feeds and so on. I think that has kind of done more harm.
DILIP: One word to describe the fintech industry today.
LALIT: I think thriving. Thriving.
DILIP: Yeah. Even I think so.
LALIT: Yeah. Thanks for UPI and DPI and everything.
DILIP: One advice you want to give to engineers of tomorrow? I think you spoke a lot.
LALIT: Yeah. I already spoke like be curious, create something and don't worry.
DILIP: One key learning of, you know, while your journey of building Groww to this level, what it is?
LALIT: I think two learnings I would say. One is like customer is everything in business. Like, you know, customer is everything. And second is you need good folks to work with. You should just feel like getting up in the morning and you should feel like, oh, I would go and talk to this person.
DILIP: Talk to this person.
LALIT: So, magic happens only if kind of you have those kind of folks in your life.
DILIP: You know, you’ve built Groww. You're a very successful entrepreneur. What keeps you excited about the future of this whole tech evolution and other (things)?
LALIT: So, I think what excites me most is that if you look at the graph of tech, how tech is impacting the world. Tech is eating everything right in all industries and where it is going. That gives me a lot of comfort that I'm a techie and there will be a lot of inefficiencies in the world. There will be a lot of growth that will happen in the world and so on. And tech will drive that. And I see myself kind of finding those opportunities to continue even in Groww, like, you know, continue finding opportunities where I can create some difference right in this in this world as a techie.
DILIP: Keep thinking that what is the impact you're making on a continuous basis? It's not just one day, one week, one year.
LALIT: Yeah, it's compounding.
DILIP: Thank you so much for spending time with us. And we at NPCI are grateful to you. LALIT: I am grateful.
DILIP: Candid conversation, open conversation. And I'm sure the engineers watching this podcast would learn a lot.
LALIT: I am actually grateful not just for inviting me in this podcast, but overall, like what NPCI has done for startups like Groww. I think I've been very candid and very honest. Like I don't think we would have existed without DPI and NPCI.
DILIP: I think it's not about one organisation, right? I think the beauty of this whole story is ecosystem, right? You know, we have been able to create such a great ecosystem, financials. And then thanks to the government, RBI to kind of allow us to build this right in a very responsible manner. But I think the responsible innovation world is kind of back.
LALIT: Responsible and inclusive also.
DILIP: Inclusive as well. I'm just saying it's so fantastic. And I cannot thank them enough to kind of enable this. They are in the policymaking side, you know, for them to take a risk, you know, support innovation. It must be so difficult. But still, I think as a country, you know, the Prime Minister, the government officials, the policymakers, regulators are really, you know, kind of enabling us to build for the future. I think that's a greatness.
LALIT: Yeah.
DILIP: Thank you again.
LALIT: Thank you so much.
DILIP: Lalit, thank you so much for spending time with us. I'm sure it's going to inspire millions of engineers who are studying or who are aspiring to be engineers themselves. I think it's going to be a great opportunity for them to learn and grow from the experiences that you shared with us. For any questions, comments you have about the podcast or if you want us to ask any questions to our guests on this podcast, please write back to us.