Software Misadventures - Kailash Nadh - On being an absurdist and building the tech team at Zerodha, India's largest stock broker - #18

Episode Date: February 25, 2022

Kailash is the CTO at Zerodha, the largest stock broker in India. In this conversation, we speak with him about absurdism - a philosophy that guides his personal and professional worldview. We discuss how he built Zerodha’s tech team, their team culture and how the team operates so efficiently while being so lean. We also discuss why Zerodha self-hosts all of their tech stack, what they look for when hiring engineers and how their systems scaled when the user base grew from 2 to 8 million in 18 months.

Let's say you are at 10 million in revenue this year. Why should you go to grow to 30 million to prove what to whom, right? Generally to investors, but we don't have any investors. So this freedom of doing things with common sense, being grounded, doing things the right way, doing things slowly and sensibly. In no rush to become number one or big or large has really shaped how we've built Zerodha and how we've grown. So this burn money, you know, fail fast, we just don't get it. We don't agree, we don't get it. And we've never followed that philosophy.

Welcome to the Software Misadventures podcast. We are your hosts, Ronak and Gwan. As engineers, we are interested in not just the technologies, but the people and the stories behind them. So on this show, we try to scratch our own edge by sitting down with engineers, founders, and investors to chat about their path, lessons they've learned, and of course, the misadventures along the way.

Hey everyone, this is Ronak here. Our guest in this episode is Kailash Nath. Kailash is the CTO at Zerodha, where he started the tech team back in 2013. Zerodha is the largest stockbroker in India, processing 15-20% of all trades. At peak, this is 15 million plus transactions daily, supporting about 9 million users. When it comes to total volume of transactions, Zerodha is, if not the largest, then one of the largest stockbrokers in the world. Zerodha is completely self-bootstrapped, has raised zero external capital, and has been profitable for years. What's further impressive is that their tech team is only 33 engineers. I first learned about Zerodha's unconventional tech team in a blog

Kalash published back in 2020 and I have been wanting to talk to him since. In this episode, we talked to him about absurdism, a philosophy that guides his personal and professional worldview. We discussed how he built Zerodha's tech team, the team culture and how the team operates so efficiently while being so lean. We discussed why Zerodha self-forces all of their tech stack and what they look for when hiring engineers. We spend the first part of the conversation discussing Kailash's early days and the events that greatly influenced his life. If you are more interested in recent topics about Zerodha, feel free to jump ahead past the 22 minutes mark. We had a great time talking to Kailash. Please enjoy this fun conversation

with Kailash Nad. Kailash, super excited to have you with us today. Welcome to the show. Thank you. And thanks for inviting me. Of course, pleasure. So we thought we would ask, start with asking you about a number that has shown up on a few blogs that you've written and this number keeps repeating itself it's 97.42 percent uh can you tell us where this number comes from it's just a made-up ironic funny number which is meant to indicate a high degree of probability that's all i could have said 99 but everyone says%. So I made it specifically 97.2, neither here nor there. It makes people think. It's just a placeholder for a high degree of probability.

Nice. Well, it certainly achieved its goal. It certainly made us think about what it was. So we went through your blog, and on your website, you mentioned that you identify as an absurdistist and we were wondering what that exactly means. Can you elaborate on that? definitely at work. So absurdism in at least my flavor, my understanding of it, it refers to the inherent lack of meaning in our lives and our lives are really shaped by extremely chaotic seemingly random events and you've got the butterfly effect kicking in, you know, changing entire timelines for you. It could have just been that one phone call

that changed your life forever or that one book that you read. And these are things that you don't plan. So my life, my personal life has been heavily influenced by the teeniest, tiniest random events that I never thought twice about, you know, completely changed my life forever. At some point, I started reflecting and realized that, you know,

I never really had any specific goals. But even if I had any specific goals, they'd never have been 100% because my life has been completely upended, thankfully, in a very positive manner by these really small events. So once that clicked at some point, maybe in my late teens, it really changed how I think about everything. And one of the benefits I found of that line of thinking when it comes to workers,

it helps you overcome FOMO and it lets you disregard status quo completely sometimes and you know you could be right you could be wrong but you get that ability to cite a really simple concrete example at zeroda we are a stockbroker now when we started building kite our trading platform this was in 2014 all trading platforms in the industry were black They had 20 by 20 grids of flashing numbers. And that's what Proos used. So I looked at it and I just couldn't grasp it. I figured, I thought to myself that it's not possible for a human being to process

20 by 20 grid of flashing numbers, numbers that flash every second, multiple times. And it just makes no sense whatsoever. So when we did the first design draft of Kite, there were just two numbers, just a two column table, just the name and the price, the name of an instrument stock and the price and the whole platform was white. I think it was probably the first white trading platform UI in India, probably elsewhere also. And everyone who looked at it said, what are you doing? Nobody's going to use a white trading platform. Nobody's going to use a trading platform that doesn't show 95% of the numbers that others show.

But, you know, something like that could have been a big deterrent and that could have really changed the timeline. But it just didn't make sense to me. And, you know, this whole absurdist worldview stuff is absurd trading platforms are black because you know someone back in the day made it black for historical legacy reasons and it's just a status quo that continues so uh it didn't bother me at all and we went ahead and that really changed everything for us so it's you So it's countless incidents like this in personal life and in my work that has eventually turned me into a full-blown absurdist from a newbie absurdist.

So it's really about questioning everything, serious quote, and just arriving at decisions objectively. It doesn't mean that you can go bonkers and be all nihilist about absolutely everything. You have to operate within a framework of reason and ethics. But it gives you that ability to see past a lot of buzz everywhere. So that's really why I identify myself as an absurdist. And I kind of wear it on my sleeve, at least on my website, because that worldview is really the answer to a lot of questions

generally about my work or what I do. It's super interesting. What are some of the cons? Because you mentioned a lot of the pros is it lets you see the world in a different view. But then when things don't have as much meaning or like you know there's no sort of purpose or you know a lot of pre kind of labels that we put on things right like how and it also like the boundary idea that you talked about or that you can't go like too crazy like out of the so so it's very much it's judgment based right like what are some cons like why isn't everybody a like is being an absurdist like for everybody uh i think i don't know if it's

scalable you know to the broader humanity really it's just there are countless philosophies right i'm pretty sure everybody's an absurdist to a certain extent. Sometimes you just have to let go of the status quo and take gut-based decisions. Of course these decisions do go wrong. So the cons are that you know stuff goes wrong but stuff can go wrong whenever. You can follow the status quo and it will go wrong. So I don't think things going wrong is specific to this world view that just hinges on the probability of success of a certain decision. If your decision, irrespective of how it was arrived at, does not really have a strong objective basis, there's a high likelihood that it might just fail. So the cons are really just the cons of any decision you take in life with work or whatever. And maybe on a slightly more philosophical level, it could be that

if you tend to question everything and if you act out on that, you know that this whole thing is fake, so I don't want to participate, you might end up cutting yourself out of social life in general, humanity. So even if you realize that certain things are just completely baseless and meaningless and people do it because people have always done it, you might still have to knowingly compromise and participate to participate in civilization. So I would say these aren't exactly cons these are just side effects of such a worldview but you can deal with you just have to be pragmatic you have to be

pragmatic about everything about this also thanks for sharing that it reminds me of uh the book by you will know harari sapiens i think he mentioned some of this and that. And I was like, well, what is anything that we believe in? Call it money, religion, whatever. It's just made up beliefs we all tend to agree on and civilization. Absolutely. If you did not believe in collective myths, there'd be no humanity or civilization or society. Well, so coming back a little, back paddling uh since you mentioned there have

been an absurd number of events that have influenced your life uh and looking back in a very positive way if you had to pick two events that greatly influenced or changed the course of your life from that point on uh would you be able to share those with us? Yeah, sure. There are three distinct events that are always fresh in my head. After I'd finished high school and I'd started dabbling in programming after completely accidentally discovering the idea of code. This was, I think, maybe in 2000 or 2001. In fact, that

was actually another event. I think I only knew two other people who had computers back then and this was, you know, southern India 22 years ago. So it was a very different time. Very few people had computers and we had dial-up internet, very few of us. And i can't recollect how it happened but must have been editing the ini file of some game to change the physics or you know some sort of tweaking that's when it clicked that oh you can change text in a place and it changes the behavior of software that that's really how it happened wait wait that's a that's a euphemism for cheating right you were trying to cheat i think it was fifa fifa 98 you change gravity to i don't know 100 you kick the ball it goes into outer space and never comes back you know some feeling around something like that but yeah

i was trying to cheat on a locally installed demo game there was no multiplayer it was not a yeah just a local install so that that's i think that's when it really clicked then i started saving web pages and editing text to change the behavior and really it really started off like that so that one tiny realization of editing i think ini files that really changed my life. So I became a programmer without really understanding what programming was or what it entailed or where the future would lead me. But I became a hobbyist program. I started writing software as a hobby. So once I finished school, this was around 2005, I really wanted to continue what I was

doing, which was sitting at home and working on hobby projects. But you know, that's not really an acceptable situation in India you can't just sit at home not go to college you know immense amounts of societal pressure and whatnot things have slightly improved today I guess shout out to all the Asian parents yeah exactly so I really wanted to continue what I was doing. I knew that if I ended up doing a computer science degree in India, the curriculum would absolutely destroy me. Four years of I don't know what and I definitely would not be able to program or continue working on my hobby projects. So I refrained from doing anything and the clock was ticking. All my classmates, they'd all started joining different colleges here and there. And someone, an old school friend of mine, who had moved to the UK with his family many years ago,

he messaged me out of the blue on Yahoo Messenger, asking for some help with his computer homework. Because he remembered that back in school I used to fiddle with computers and I think I hadn't spoken to him for like five six years at that point we were all really young 15 16 years old and I helped him out and in the end he asked me you seem to be good with this stuff you know code. Have you thought about studying in one of the universities in the UK? And up until that point, the fact that it was possible for a student like me, post high school, to go abroad and study, that thought hadn't occurred whatsoever. It

just had it. But when he asked that on Yahoo Messenger, just that one line, that really changed the timeline for me. And that is when I started digging into it, researched, spoke to universities, and it worked out. So I think a few months later, I actually went to the UK to do a B.Sc. in Computer Science. I got a partial scholarship, so I ended up in the UK. That one line on Yahoo Messenger completely changed my life there. And could have worked out, could have not worked out but it did change drastically. And the next big event was, I was there for three years, did my Computer Science, you know, it went well,

the degree and again the next big life decision is to find a job or figure out what to do or study more I just wanted to sit again in my bedroom and work on my hobby projects and it just seemed so wrong that I'd go all the way to a whole different country study in a university get a degree and come back and live in my bedroom just doing my hobby projects and making a living doing freelancing. So that was a really difficult time, you know, mentally to figure out what to do next. Again, somebody, a senior student from a different batch reached out to me saying,

hey, I heard of you from this other student in that other batch. You seem to be good with programming. I need some help with my coursework. Some completely random person who I'd never met. And I helped him. And after I'd finished my degree or no, sorry. After I was just about to finish my degree, he reached out again saying, this is my final year project here. I really, really need your help. This here. I really, really need your help.

There's this Java program. I really need your help. And I was like, sorry, I don't really do much Java. I can't really help you. And I politely declined. But he was persistent. I think he messaged me every two weeks. And after like two or three months,

he messaged me again. It was around 7 p.m. in the night. It's like a flashbulb memory I remember it so distinctly you know something that changed my life and he said my final year project is due it's next week if you don't help me I'll fail and this person he's not my friend he's just a person that I've spoken to once or twice you know in a whole different context. So for whatever reason, at around seven o'clock and was really cold that night, I got on a bus and traveled an hour and a half to this person's place and sat on his personal desktop to help him out with this Java program. Just like that.

Just, I went. And it was around midnight. So he said, you know, and it was around midnight. So he said, you know what, my roommate is coming back. You can't crash here. You have to go back. I was really, really pissed off because it was around 1 a.m. in the night. Yeah, it was 1 a.m. in the night. It must have been minus 10 or minus 5 degrees.

And, you know, the bus line had shut down, that particular line. There was no way for me to get back home. And it was an hour and a half again. And I really thought, you know, all I could think of was unsavory words. But thankfully, you know, I phoned a friend. He drove all the way there. He picked me up, you know, we went back home

and it was really, really, you know, pissed off. Then he messaged this person, the student, he messaged me a week later saying, I passed, I barely passed my course you know coursework uh it went through thanks for helping me with the job you seem to oh yeah and it was a simple neural network model that was the program and he was a I think MSc student master's student and he said to me just this one line there is my my co-supervisor's professor uh sorry dr christian hike uh he's looking for grad students to help him with ai projects point to speak to just like

that casually he dropped that one line along with this you know thank you note and i sat on it for a week and i'd forgotten about it a week later later I messaged Professor Christian Heick, he was an AI researcher at the university, still is, at the Middlesex University in London. I messaged him, he asked me to go meet him, I met him and he said yes I've got funding grants for this particular project, you know, cell assemb networks simulated neural networks uh i need help with building some of these models uh if you can program why don't you help me out and i was super happy uh seemed like an interesting project and i did that so we wrote a paper and we published a paper and that worked out really well. And that project became my final year project for my B.Sc. My B.Sc., my degree ended right there. And I had to decide what to do next.

And out of the blue, professor, he's professor now, he was a doctor back then. So Dr. Kristen Haik said to me, why don't you do a PhD here? And up until that point, you know, those three letters hadn't even occurred to me, why don't you do a PhD here? And up until that point, you know, those three letters hadn't even occurred to me. I was, I didn't even know it was possible to do a PhD without doing masters. Apparently it is in the UK. If the university is willing, if they like your work and whatnot, they can fast track you, they can let you skip masters entirely. So he said that, and I thought about it and I said i thought to myself that of course it's a really interesting area of work the project was really interesting that i worked on

so i could just maybe just do that for three more years and i could continue working on my you know side projects and you know i stayed on for three more years i ended up doing a phd and see just that one line my completely ad hoc decision to get on a bus on a really cold night to go help this stranger that changed my life forever so then I ended up in the UK for six years and once I did my PhD I had better clarity but I really still wanted to come back and you know work in my work work out of my bedroom working on my own projects. I didn't want a job. But because I had obtained a PhD at that point, there were several really good offers to do postdoc research at multiple universities. There were industry offers because this was a PhD in AI and even to an extent applied AI, neurocomputation.

This was 2010. So I ruminated, thought about it really hard for like four or five months. And I decided to pack my bags and just come back to India, just like that. And I, you know, six years there, had a great time, studied a lot, amazing programs. And I just came back to, I'm from this really small state called Kerala from India, one of the smallest states. So I came back to where I left from and with the goal of sitting at home, working on my hobby projects, working on whatever I felt like working. And of course, you know, trying to make a living doing freelance programming and i'd figured that out you know i was making a decentish living so i come back to this is now the third big event i i come

back to kerala and i realized the internet really really sucks it's still at 64 kbps and as a student in London I enjoyed fast broadband internet and I immediately regretted you know that one thing slow internet that really really you know bumped me out bumped me out and it had only been a week after I'd come back to move back to India after spending 6 years away from India and my cousin said to me that Bangalore has fast internet and I just left for Bangalore. I just packed my bags and got in a bus and I came to Bangalore where I am right now. This was late 2011 or early 2012.

And Bangalore indeed did have broadband, 16, 8 Mbps, 16 Mbps. So I found a place, I rented an apartment here and it just moved to Bangalore. That whole process took like a week, 10 days. And I ended up in Bangalore because Bangalore had fast internet and my cousin told me that Bangalore had fast internet. So I moved here, you know,

set up my apartment. I started working from home, you know, freelancing, working on my projects, experimenting. And again, yet another message from a school senior that I hadn't spoken to in like maybe 10 years. He messaged me on Facebook and said, Hey, I remember you from school you used to be used to fiddle with computers I have this I have this idea to build a financial app would you be interested and I asked him where are you he said I am in Bangalore because anybody who wants to build an app ends up in Bangalore eventually so and I said I'm also in Bangalore, let's catch up. So I caught up with my school senior whom I hadn't spoken to in a decade. And he had this, he'd ended up in finance.

He was in Mumbai, the finance circuit, and he quit the finance industry because, you know, he was getting dated and he had this idea to build a really simple investment app for the masses. And this was 2012 and investing and trading was really complex here you know there was bureaucracy there was a dead red deep the fees were extremely high you know hefty fees and all of that and we started i liked the idea because i was always looking for projects to work on and we started working on it although i have zero background in finance or capital markets and he had this other friend who was also working with us and we'd go hang out in his apartment

you know maybe 20 minutes away from where I was staying Bangalore and it turned out to build a financial investment app in India you had to get your regulatory approvals and you could only do that by partnering with a regulated stock broker so you get your regulatory approvals and you could only do that by partnering with Irregulated stock broker. So you couldn't just build software and you know, take it to market. It wouldn't work. So we started looking for Partnerships with large stockbrokers. Nobody responded in the incumbents typical incumbent behavior but the second friend who's a Friend of ours who at whose place we'd hang out, Zerodha's first office, in a small little office

on the first floor of the building, happened to be across the road. And this was 2012, Zerodha only had, I don't know, maybe 2,000, 3,000 users. It wasn't an online broker. Nitin was experimenting with the low-cost pricing model. And we were looking for brokers. And we'd just walk around on the street.

And one day we spotted that board that said, you know, Zerodha, a name that we'd never heard of. And it said NSE, BSE, MCX, you know, three big Indian stock exchange. And Abid, with whom I was working on this, my friend, my school senior, he walked in one day and said hi to Nitin and introduced himself, saying,

we're building this investment app. You look like a broker. Are you a broker? Can we partner? And it just worked out. So then the three of us went, met Nitin and spoke to him. And he said, we're a super tiny broker. We have this disruptive pricing model, but we'd be happy to partner with you. We'll try and get you the, we'll try and get you the regulatory reports. This only happened because we were hanging, we were hanging around the same neighborhood

and Zerodha's office happened to be across the road. If it wasn't there, we'd never heard of, we'd, sorry, we'd have never heard of Zerodha and I'd have never met Nitin. So we spent the next nine months trying to build this app. In the end, you know, a regulator shot it down saying, no, no, we don't want any technology in this market. We like the way things are. And we disbanded, gave up building on the app.

But Nitin and I, we started getting along well on a personal level. We liked each other, liked each other's philosophies. I liked his way of looking at the business. He liked my way of looking at technology. And, you know, one day we said, said you know what the other project is over. We all went a separate place but why don't we work together and try and build stuff here in the stock brokerage industry where nobody's really building any technology and in in 2013, we started Zerodha Technology. And the last eight, nine years of my life has just been that. You know, that one blue board that we spotted across the road

that happened to be in the same neighborhood completely changed my timeline yet again. So these are the three big absurd events that I always look know look back upon and this always gives me perspective when i'm when i'm confounded with some i don't know big life puzzle i just look back and like oh yeah it was all random my life was random anyway may i school you know sorry sorry i know it was a little i went off on a tangent but oh no no no this is this is great uh there's no recipe on this podcast, so this is great.

You say it's random, but I feel like every time that happened, it was like, oh, someone contacted you because you were very good with computers and they liked you, right? Which I feel like, yeah, you know, the randomness is probably like a necessary condition, but, you know, it's definitely not a sufficiency condition, right?

I think randomness here not a sufficiency condition right where um i think randomness here is a big enough condition uh my being a programmer and uh my projects and whatnot they were of course they increased the odds significantly but if i wasn't a programmer no one ever messaged me but just because i was a programmer does not mean that you know people would message and things would turn out this way. So that randomness that, you know, that bit of serendipity is a significant two digit percent component in this, you know, holding of odds. I like that. So you're like, you're still making a lot of effort, right? Because you want to push up your probability.

But you're also, I feel like you're never disappointed because you know that the odds, know are not that great so absolutely you just summed it up so if there's no effort the probability is really low the better you are at something the more you more effort you put in your odds just go i absolutely nothing is guaranteed because you need luck chance randomness but if there's no effort there's nothing the odds are really really really low rana that should be uh you know the motto of our podcast it's like low expectations we'll try but please low expectations yeah also sounds like recipe for happiness okay so fast forwarding ahead you've been at zeroda as you mentioned for eight years now and uh you're the cTO at the company.

And Zerodha is the largest stockbroker in India at this point. To give our listeners some context, would you mind sharing some numbers around the number of user base you have, the volume of transactions or daily trades that you're doing these days? Yeah. So I think we are closing in on 9 million users here. And by users, I mean people who've done full KYCs, people who've paid a fee, people who signed 50 pages digitally signed. So yeah, actual users.

And we hit a peak, I think last week, where we processed, I think, 15 plus million stock market orders. And these are actual transactions coming from end users, sending them on, using their mobile apps and web apps. So not algorithmic transactions, but 15 million retail trades. I think that constitutes to about maybe 15 or 20% of all stock market retail order volumes in India across all stock exchanges, which is a significant percentage of all Indian stock market activity. And there are a million plus people who are always connected, live streaming, who spend way too much time looking at the numbers.

And I think last time I checked, we were streaming maybe 50, 60 million packets of stock market ticks every second to so many users. And at market opens or when news breaks, we could we see tens of thousands of orders, actual stock purchases and sales coming in at 10 or 15 or 20,000 per second at that rate. So the volumes are really high and I would think that we're probably the most active retail stockbroker, maybe in the world, when you look at retail activity. Not really the value of transactions because, you know, Indian rupee versus USD, you can't really compare. But the volume of transactions placed by actual end users interacting with stock markets, it's, yeah, we're probably the biggest, if not one of the biggest i say that because last year when charles schwab in the u.s

had posted a press release saying we just hit a record high of end user transactions and we crossed two and a half million transactions and at that point we were at seven or eight million so yeah that those are some numbers oh that's i mean that scale is incredible and i think what's more is that zerodas from my understanding has never advertised not spent anything on marketing like all this growth has been organic where users are actually paying a commission to sign up they are signing up to actually do trades and beyond that i think what's amazing is that you have a tech team which you started as a solo engineer in 2013 and today it's still approximate i think around 30 people if that's correct uh maybe 32 33

something like that and so 33 of you and you've built as you mentioned based on the numbers at least like largest stock broker if not the largest one of the largest in the world in terms of volume of transactions so for any engineer or anyone actually listening to this it's mind-blowing that wow and to be honest the first time when i read your blog where you mentioned this uh so one of my friends texted me this blog saying hey you have to read this. This is awesome. And this friend is in Bangalore. And I went through the blog as like, holy cow, this is amazing. Like one, I really like how you wrote the blog and just understanding the philosophies of how the team operates. And we want to dig into some of that. But starting with like one thing, which I want to begin with is so you mentioned

how you've scaled the platform uh and we'll link all of your blogs and the show notes and i encourage all our listeners to actually go and read because entertainment aside there are really good engineering lessons there and one of the things you mentioned as your obvious secret sauce is self-hosting, which might not be that obvious if you ask many people. So I want to ask, can you share what aspects of the stack when you self-host so that we have the context to build upon? We try to self-host everything that is practically possible. So our trading platforms, the databases, be it an RDBMS or be it Redis or Clickhouse,

really large installations, they're all self-hosted and self-managed. And these are the components, but everything really about the business, our HR management system, employee portal, support ticketing system, the CRM, all sorts of systems, we've self-hosted them.

So we don't really use any SaaS CRMs or support systems or business tools or any of those things. Everything is self-hosted. If it can be self-hosted where the trade-offs are worthwhile, it's not impossible to manage, we self-host it. And so far that has worked for 97.5% of everything.

So it has worked really well. one question that i constantly get asked is but who manages all of this you know how can you manage hundreds and hundreds of systems i think it's a myth that many of these systems need active management the support ticketing system we use post tickets ancient piece of software thousand plus people log in every day use it we've been running it for six years. Every year we log in and once we were at a certain point where we've accumulated millions of tickets, we do some basic database cleanup and it just runs forever. Some of these tools, they're so rock solid, they will just run forever. And I think that fear that oh the system will need maintenance for who

will look after this I think that that really holds back a lot of organizations and stacks and if you're an absurdist you really don't care about that see this group once you objectively understand that that this is good software and you can run it for a long time. You may as well just do it. So yeah, we self-force absolutely everything. Technical backend systems, you know, components to business dashboards and systems that our non-technical employees use. Has there been any open source software

where you've tried it for a while and then it just doesn't work super well and then maybe you compare a few options, but then in the end, there's just still that maintenance cost to it such that't work super well. And then if you compare a few options, but then in the end, there's just still that maintenance cost to it such that you have to switch. Like, has that ever happened? Yes, it does keep happening,

but it hasn't really happened at a disaster scale. Yeah. I'm trying to think of an example here. So we used to use PHP List. It's an old, quite popular, it's a mailing list manager. We started running into huge performance bottlenecks, evaluated two, three, four newer systems. Everything was worse.

We came back to PHP List and it just wouldn't work. In the end, it became unviable and the cost of hosting millions of users on an external mailing list platform, let's say Mailchimp would have been quite a bit. And we ran into this bottleneck three, four years ago. Now, I don't really know if this is the right way of looking at it, but throwing money at a problem is always the last option that we consider in our tech team. That's how we are looking at things.

We try to self-host or we try to maybe buy something. We never really had to buy anything. We'll try to roll our own. If none of these things work, we're like, oh, let's throw some money at it. So that's how I ended up writing ListMonk. It's a personal project of mine. It's an open source newsletter, mailing list manager. So because none of these systems worked

and because I really did not want to upload all our customer information onto some SaaS system out there, I ended up writing this as a prototype, but it turned out to be a good project. It's scalable. It works. It's stable. We use it internally.

Now, lots of people use it externally also. So that's one example I can cite where we exhausted all possible open source options and we had to invent our own. So this is super interesting, right? Like the trade-off part that you mentioned. One conversation I can imagine in my head that some teams would have is

we need a mailing list software. The previous option isn't working that well. Either we roll our own or we pay a vendor like MailChimp or someone else. And people usually think about the dollar cost that they would have to pay versus, well, there is engineering cost on top of that there is, well, for many people or many teams, opportunity cost. That one, we need to find someone who will do this. And if that person does this, then what about all the other things that we need to get

done? That's the trade-off people think about. so the amount of money in general is like engineering costs as well as that opportunity part how do you go about thinking about these things in general and they is it default that we'll sell first because we've done the math or that math is something that you always look into every time you have to do this? We do consider all these trade-offs that you just mentioned. These trade-offs go into every single self-hosting decision. So as I said, just because we like to self-host doesn't mean we are doing it without an objective basis. If it doesn't make sense to self-host

or if it's impractical or if it's not good enough, we wouldn't do it. But like I said, so far, self-hosting has just really worked out for us in all aspects. And the thing about money cost is that, yes, we could have just thrown some money at MailChimp and this was three, four years ago,

but last year we added 6 million customers you know the whole 2020 you know sorry a year and a half ago the whole explosion happened and nobody knew nobody in the industry predicted that the industry would grow manifold in a year so if we had computed costs based on our the current based on the state of the industry at that point you would have been vastly wrong now this is just one mailing the mailing list manager is just one decision right there are hundreds if not thousands of such decisions and the dollar cost eventually does add up so it might not it might be immaterial for this one decision or the next decision or the next 10 decisions but over many years they all add up so one of the reasons uh zerodha is a profitable you know we are a low cost but high margin high profit margin business and we've never had to raise funds is because of these early decisions you know uh because of the frugality component that we incorporated into all our tech and product decisions.

So the amount of money today we save because we self-host is immense. It is a huge component of the entire company's profit margin. So these things definitely add up. And on that note, if you look at our net infrastructure costs, you know, VMs, AWS, you know, racks at different data centers, etc. It's peanuts.

If you look at it as a percentage, it's laughably low. And that again throws out the misconception that to scale a percentage of all your income has to go into infrastructure. That's not really how software works. You know, software doesn't scale linearly. So all of these decisions as a framework, they make us what we are today. And they always pay off over a longer period of time. And like you mentioned, this is very contrarian, right? Because in Silicon Valley, the mentality is burn money, right? And then try to grow as fast as you can.

And when you guys were maybe at the earlier days and thinking about, okay, you know, what does this look like in five years? Was not raising money always part of the plan like from the beginning or how did you guys deal with the temptation of okay let's go raise some you know vc money and then we don't have to worry about these sort of things uh thankfully this is where our philosophies align one of the reasons i nathan and i nathan is the founder and ceo of zerodha. He founded Zerodha in 2010. 2012 is when we met. And this is where we really hit it off. I absolutely like his philosophy,

his super pragmatic, super practical, grounded philosophies on how to run a business. So at Zerodha, we've never had any specific goals whatsoever. We never said we need to get 100,000 users or X revenue or a million users. When we got 2 million users, we never said, right, what do we do to get the next 2 million users?

We've never ever done that. We've never had a projection. We've never had any of those things. Our philosophy has been really simple. If you're doing technology, we'll do technology the objective technology way. If we offer financial services, we'll offer common sense financial services that make sense for the user. So we want these services to be really sensible to the user and of course we want to make profits and be sustainable. So you put all of this together, we really took it slow.

And we grew very organically. We were absolutely never in a rush to get to any arbitrary goalposts. One million users is a completely arbitrary number. Why not 800,000 users or 5 million users or 100 million revenue? You know, these are all completely, completely arbitrary projections, right? So-called projections. They're so baseless. Why should you, let's say you are at 10 million in revenue this year.

Why should you grow to 30 million to prove what to whom, right? Generally to investors, but we don't have any investors. So this freedom of doing things with common sense, being grounded, doing things the right way, doing things slowly and sensibly, in no rush to become number one or big or large, has really shaped how we've built Zerodha and how we've grown. So this burn money, you know, fail fast, we just don't get it. We don fail fast we just don't get it we don't agree we don't get it and we've never followed that philosophy now i'm not implying that our philosophies work they have just worked for us that's all they may work for others too maybe this is a horrible question but like it's hard to be contrarian right because you look at all these

other examples um especially successful ones right where they are raising a ton of money and then they're getting crazy evaluations and the things seem to be working really well right and then you're looking around and then you don't have that many other companies doing what you guys are doing right how do you stay true to yourself yeah this sounds really cliche ronak do you have a better is there a better way to phrase this like is it just not giving a crap about what else is going on outside or like do you do you have those like moments where you're like oof like is this the right you know i mean at the end of the day you're human human. You do have feelings, right? So once in a while, you think of these things. Is this right? Right, right. How do you overcome those moments of weakness? Is that what they call it?

Because if you're an absurdist, you really know that there is no right or wrong. So you use the word contrarian, but I have to be contrarian here and I disagree. So your contrarian, your view is contrarian to another view. If that view is, let's say objective. Right. So, you know,

burning lots of money, growing valuations. It's a trend. It's a fairly recent trend. Businesses have existed since the dawn of civilization. This whole, you know, burn fast, grow valuations is maybe 20-30 whatever years old. So it's just a thing that people do. Yes, it seems to be it is the status quo right now, but it's just status quo, right?

Who said it is the way to do it and what is really even the point of all of this? So you might feel a little left out if you're aiming for, let's say, 10 billion valuation. But let's say your company is valued at 2 or 3 billion. The fact that we use billions as a base holder today is really telling of times. But whatever, X. If you don't really care about these arbitrary numbers or valuations then none of these things really matter you don't feel left out because you're not really aiming for that in the first place so we the group of people at Zerodha we have a vision set very clearly we like doing what we're doing

and I enjoy that I don't really get stock markets at all i still don't understand it i really like working in the team that we've built you know it's fun engineering challenges product challenges and the and the business folks that we work with uh nathan it's all common sense and sensible we make a good living uh there's a sustainable profitable company with like a 10-year cash reserve now with zero external funding we don't have to prove anything to anyone no investors to answer to it's fun it's like a hobby so when you can make a really good living doing your own hobby and you don't have to worry about tomorrow or the next 10 years when it comes to you know financial sustenance that's a really good place to be right so if you're valued at 4 billion 5 billion 10 billion it really doesn't matter

you you really only question yourself or feel left out if you know these arbitrary goalposts are your goalposts but you don't as i said we've never had any goalposts going with the flow and enjoying it i know it sounds cliche but no i i think that's incredible uh not feeling that FOMO as much or if and when you do being able to kind of put it aside and go back to that well the world is absurd let's just keep doing what we're doing um exactly yeah i think i would call that a superpower uh for for many people many people would want that. So in terms of going back to self-hosting a little, you mentioned that you self-host

and this is something that has worked very well for you. It has saved costs for the business as well. So the trade-off is well worth it. But as someone is trying to self-host any software, and of course, software is not perfect. It breaks for reasons out of our control. Machines go down and cables get cut underwater someone goes digs the street and hey we lost connectivity

it has happened for any person who has been running software but this is not something that start in school like in school you learn about a programming language you learn how to write code but you don't learn software engineering necessarily you don't learn the operations part of it and i heard you uh you were saying in a different conversation where you mentioned that you hire a lot of folks from school who are just graduating so how have you been able to build that operational muscle as a team for folks who have never practically run software at scale we don't really have a framework and there are no formulas. It's just happened. Our team has happened to evolve a certain way.

And I guess we had the luxury of trying to do, to learn by trial and error. We were not burning cash or acquiring large number of users or, you know, hinging on evaluations to prove anything to anyone, right? So I think that has played a key role. So you can, you know, build things slowly. You can make mistakes, contained mistakes,

and learn from them. So all the lessons that we've learned, we've learned gradually over a period of nine years. It's only the last three, four years where our scale has just increased exponentially. Until then the growth was really slow and linear and our growth as a stockbroker is very closely tied to the performance of stock markets and stock markets were you know rather flat for many years throughout 2010s so it's it's just opportune if you were in a very different environment where you know there was stock market volatility every single day maybe things would have turned out to be different i don't really know but what has worked for us is doing things by trial and error

and of course it's one key thing. One key aspect is that it's not just philosophies. You need actual technical competence to self-force and manage these systems. If somebody lacks technical competence or expertise, they should not self-force. I mean, in a critical business environment. So you need that. That's a given. You need to know what you're doing.

Your decisions have to be technical and objective. So thankfully, we had those in place. I had, as a hobbyist programmer who'd freelance for over a decade, I had good enough experience in building and maintaining systems for others and user expectations and whatnot. So as we hired and as the team grew, we did, the business was also going slowly and organically. So we really have the luxury of time to self-host,

to break something and make it again, learn a valuable lesson and so on. And then when the markets, when the industry, the environment really just went through the roof, we'd already built like a solid technical framework, foundation in the organization and expertise. So, yeah, I'd say the luxury of time really was an outcome of luck and the early 2010s that we got to enjoy early to mid 2000s yeah it's amazing you had the luxury of time to learn these lessons i think that's a an engineer's dream to be honest like you take a system break it and see how it works and learn from that now you mentioned that this is something that folks learn on the job and you still have to hire for competence.

So, by the way, I checked your careers website and there are zero job openings, which is pretty unique, to be honest, and fascinating as well. And one of my friends, this was a few months back, who lives in the US, and he pinged me saying that, hey, and we talk about Zerodha and he's like, hey, Zerodha doesn zero that doesn't have any openings and i was like that's pretty neat and that's still the case right now so how did you find these folks and how do you how do you evaluate that technical competence to know that okay this person's gonna be a good fit and they will know what they're doing and have the ability to learn on the job and also the intellectual curiosity right because to be able to break these things down and then build it up again you you want people that are actually really interested in how things work right versus i'm very lazy you know it's like oh maybe we should do something else here you know but yeah how did you go without doing that uh it's really looking for these traits as you just described.

There are certain traits, curiosity and I think hobby programming is a big indicator. If you are a hobby software developer, you write hobby software because you enjoy doing it. When nobody is really watching you or judging you, you are not doing it for your employer. You are doing it in your free time because you like doing it. And that's a huge indicator. And so our hiring has been, again, very ad hoc, two people a year generally on average. And I found all the engineers on random forums or local forums, really. Bangalocentric job boards, forums.

When I say job board, I don't mean like a big job site, but a job board where people only post like 30 job openings a month because it's so niche. And that's how the connection was made to potential applicants and I've spoken to lots and lots of people, countless people and 97.42% of the people don't cut it. Just like I'm guessing that's true for every industry. Most applicants, they just don't cut it.

And when you speak to someone, you can kind of figure out their attitude towards work their expectations from life in general and you know that that a very open non-recruiter conversation with no agenda just to understand the person goes a long way that works when you're only trying to hire one two people here not if you're trying to hire you know hundreds of people a year so yeah so it's really been about conversations and one of the things that I've done always is to set the expectations straight you know very clear you know we're not changing the world we're not going to you know empower next billion Indian people with

financial stuff you know I tell people that most of the time you'll be doing really boring stuff, you know, processing CSVs for the industry. The regulatory risk is insane. You might go to jail for a mistake. So I think it's very important to set expectations very clearly. We've never sold a vision to anyone.

And so people who get it despite all of this, despite the groundedness of the whole thing, I think they tend to be of the right wavelength because they're not coming in to be SDE level 6 tomorrow because there are no SDE levels at Zerodha. Everyone's a developer and that's that. And everybody has to do whatever has to be done. So these things are laid out straight in

these early conversations. And that's a great filter. People who can't really deal with that, they just they don't follow through. And then give someone a really simple, super simple task. I tend to give

a Python or a Go task and really simple you know make this HTTP call, fetch this zip file, extract the CSV, process it in a certain way and we're looking at really beginner levels of beginner levels of full stack software development and from this really simple task, 300 of 100 line tasks, you can tell a lot, how someone structures classes, how someone writes their functions, how someone injects something, how someone invokes something, how someone writes the readme. These are all flags and you take all those flags in you get a you know certain intuition about the person yes this person he may not know a lot now but this person he or

she may will pick up quite a bit or they will learn fast and this has worked out really well for us so I'd say two people have left the tech team for personal reasons, not really for reasons related to career. We've had to let go of four or five people over the last nine years because they just didn't cut it. But for the rest, these things have worked. We work really well as a team.

There's no politics. There's no reservations. Everybody's chill. Everybody's beaming everybody else and everybody has to do whatever has to be done nobody can say I don't like CSVs, this is so boring if you have to process a CSV

you have to process a CSV I do it, all of us do it so it's about setting expectations clear the curiosity piece I feel is really important here and from my experience where I've seen, maybe not the con, but if you take it to extreme, right, you have people that really pursue intellectual curiosity, right? And then they don't, right, it's kind of like when you ask them to do the CS piece, like, let me come up with a neural network, you know, maybe 20 layers.

Let's put a CNN on top of it um yeah how do you filter like i guess so you just set expectations with them in terms of but then do you ever get people that like you know come in they're like oh yeah yeah sure they'll do it but then um you know how do you manage that um it's like the right amount of curiosity, right? I think, so even though we don't really have a code of conduct, there are certain general philosophies, principles that the entire tech team works on. Unwritten rules that everybody knows. So one of those things is that you don't do impractical things.

And how to define something is impractical. If somebody is deeply curious about something, you can go to great lengths to experiment with it with no tangible benefits but at the end of the day we are a heavily regulated financial business and our work carries a lot of risk and everything you know is we are held accountable by multiple regulators. So that sort of work environment and the general principles with which we work, it kind of has created a culture where people don't really, you know,

completely go off the tangent. Yes, all of us experiment, but we get together and discuss and debate objectively. If many people say that, no, this just does not seem you know worth it then others act accept so the question really comes down to that you know about egos if you're deeply convinced about this one thing and you really want to do it and and the others in your team are saying

don't do it you might still go ahead and do it now that's ego politics thankfully you may or may not like the decision in the team but you have to accept it it's about mutual respect so that those two things that we have to have mutual respect for the team's decisions and that we'll arrive at decisions by debating them objectively and as scientifically as possible you know we sometimes write down bullet points pros one two three four five, three, four, five, six. Cons, one, two, three, four, five, six. You know, pros outweigh cons.

So let's do it or not do it. Even if you don't like it, even if your gut says otherwise, after enough deliberation, you have to go by the objective consensus. This works. This has worked really well for us. So the ego piece, like, would that be like a red flag during the interview if you feel like this person would have too much ego piece like would that be like a red flag during the interview if you feel like this person would have too much ego like um coming in absolutely absolutely so uh one of those

other things is you know technical skills technical ability sheer technical ability versus empathy for the team i would generally i've seen that at least in the industry here, you know, companies that I interact with, people, people way share technical ability way more than, you know, empathy. So to have a team that works well, you need the right balance of technical skills and, you know, empathy with the team. If there is someone who's supremely skilled, like a 100x programmer, but if they can't work well with the rest of the team, that's a no-go.

I mean, that doesn't work for us at all. I don't know if it may work for other teams. That doesn't work for us. It doesn't matter how technically skilled you are you have to do whatever has to be done within this team and you have to work with everybody else and you know people don't have the same skill levels right even within a three-member team the distribution of skill and talent will be very different so this this whole thing about, oh, he's a better programmer than she is. So those have to be weeded out very early on.

That's why we don't have structures or levels or hierarchies of designations that imply that yes, this person is a better programmer than this other person. Everybody implicitly knows that certain other people have their strengths, certain other people have their weaknesses, with within the team we have to all work together and do what what has to be done so setting that very clear on day one that you have to continue ego and work with others that yeah that's absolutely paramount that's how we hired and built the entire team yeah no the empathy i think also totally agrees huge and my first personally my favorite interview question is like a code review question where you present a code review to a candidate where then they give

feedback to that code review so sometimes you would have candidates who are very talented on the code picking like picking out all the mistake but then the way they would give that feedback right sometimes it becomes kind of clear where they just can't really put themselves into other people's shoes but do you guys have any sort of specific methods to like to select for empathy because that's pretty soft or like pretty vague right uh it's very difficult it's intuition when you speak to someone when you speak to someone about you know their life

their perspective their work how they write software you know about technical things I think it kind of emerges of course it's just a guess but it is a largely

informed guess and as I said that has so far worked out well for us but it's impossible to quantify people's empathy into a score

I mean any sort of recruitment system that attempts to do that is flawed you can't turn humans into numbers yeah are there specific questions that have helped you get more signal on not just empathy but some of these so-called soft skills in terms of how the person would fit in the team, not enough ego, good empathy, things like that? No specific questions. These conversations are never really interview questions. They're just free-flowing conversations, like the conversation that we're having right now. So pretty much all of these conversations

have gone exactly like this. You start off with something. Hey, that project that you did, that's nice. Tell me about it. And it might just branch off into entirely different things. So open, casual, free-flowing conversation. As I said, it's not really a scalable recruitment methodology, but if you're only trying to hire one or two people a year,

it's possible. But hey, props to Ranak for looking out for our listeners who are interviewing for jobs, trying to seek out some questions. I respect that. Thanks, Rana. I was rather trying to figure out as an interviewer how I could do a better job of getting that signal. But hey, it worked both ways. Smooth, smooth. So there are a couple of things you mentioned, which I want to come back to. One of them was that you don't have these levels defined in the company like SD123.

Many companies do that. Well, I should say most do that. So in your case, how do you go about providing feedback? How do you go about performance and growth included in that team? Because at the end of the year, people are expecting to know what they're doing well. And money. Yeah, but the thing is also what they're being evaluated against like what's the expectation are they meeting that or not so how do you go about that with when you

don't have these levels defined i think the word i'd like to use is holistic like a holistic approach but i understand the not so scientific connotations of that word but I can't really think of any other word so it's really there's no metric there's no system there's no matrix there are no measurements everyone in the team they have grown to own projects own responsibilities and you really evaluate them against the ownership they've taken. Somebody who's working on the trading platform, it's not possible to objectively evaluate their technical work

or product work with someone who's working on an internal system only for the business folks to look at. How could you possibly compare, right? So it's not right. So you evaluate people against the ownership and responsibility that they've taken up. This person has taken up this responsibility. How well have they done it? And that's key. That's really the primary prospect of evaluation. The other one is yes this person this developer

has taken up this responsibility to run it well but have they helped others or have they spawned of other projects have they helped other projects have they you know found time to spot issues in other systems and help them some people are really really keen, right? They like to be involved in lots of things. They go help lots of projects, lots of people. Some people are just content with just doing one thing, which is also fair. You need a mix of everything. If everyone is content, there's no innovation. If nobody is content, there's no stability. So it really depends on that, the ownership, sense of ownership, and the quality of work they've done in the projects they own, the criticality of these projects.

Some projects are insanely, insanely risky. Some projects are not risky. And the kind of work you've done on risky projects and the kind of work you've done on risky projects and the kind of you've done work you've done on not so risky projects and what new have you done this year is it just the same thing as last year which is okay i mean like i said some systems just need maintenance forever but have you attempted anything new have to have you tried to learn something new and have you presented that to others saying hey i learned this new thing you know maybe you have to have you tried to learn something new and have you presented that

to others saying hey i learned this new thing you know maybe you can try this you know all of these signals put together gives you a pretty good picture of what everyone's doing thinking how people are working and because it's such a small team everybody knows what everybody else is working on everybody speaks one-on-one with everybody else so it's i have a good grasp of what everyone's doing in the team at any given point not because a micromanager monitor we don't do that either but because i am somehow involved in their projects you know via random. This tiny bit of this system also overlaps with their other system. So I end up working with them and I get to know what they're working on.

So I guess this is only really possible in a really small, tight knit, flat team. So this whole method of evaluation, feedback, etc. is based on a deeply personal and hands-on understanding of each person and their work

they are only possible in a small team i guess so it's pretty unique i would say like everything else that we have talked about as you mentioned that there are people who have started taking responsibilities for different systems and what it sounds like is you have a team who can perform really well in a very autonomous way. And people act like owners. They take up responsibility and have initiatives. And curiosity is something that you look for. So in an environment like that, how do you figure out who works on what in a given time frame? Like, okay, you've got to plan for something.

Either that's a fix that you're doing. Either that's a new feature you need to build. How do you balance that part versus that there are a few people who own that system whereas someone else who's just curious about that. How do you balance who works on what in that case? I think that also is decided really on a case-to-case basis, but there is some foundation there. If somebody owns a critical project,

they are in the best position to continue working on it. Now, just a hypothetical example, our trading platform, it's an extremely critical front-facing system, right? The amount of risk is immense. And let's say there's a new feature that has been proposed, but we have a long list of critical features that the principal developers are working on. And this is other feature which can be done slowly and is not as critical. Now that is ripe for being given out to someone

who's interested in tackling it. But if there's a critical change, the principal developers have to do it. 99, 97%. It's just objective. It's a very objective decision. The criticality, risk, the necessity of a certain

change or feature and there are some projects that come up in a small projects that are completely unrelated to anything they're like who wants to do this and if nobody's willing to do this they're like you know what you do this so or i'll do it myself so it's that it's on it's it's a very it's on a case by case basis really and everything is every every decision is evaluated objectively. It feels very human. Like, instead of systems and rules, you know,

and it's like having good judgment, it sounds nice. I was going to ask, like, humans are non-linear. Systems could be linear. So, when you have smart engineers in a room they have strong opinions some of them hold them loosely some of them not so much and people can have bad days with absurdities in the world so how do you go about resolving conflicts when uh when you see them and there's

consensus is not being reached either in a timely fashion or in a way you would like uh i have very strong opinions many of us in the team have very strong opinions but as i said i think it really comes down to mutual respect. Even if I have a strong opinion and we hit a deadlock, who makes that 0.1% of extra cents, it really goes to them. It could be me, it could be somebody else. So it's never 50-50. It's always 49.99 and 50.01. So we have enough mutual respect to just lean towards that.

And thankfully, because of this, we rarely have conflicts, especially with technical matters, right? It's not, I mean, we could debate on a piece of art saying, this is nice for such and such subjective reasons, and we could be deadlocked forever. But when you're evaluating, let's say, a Postgres database versus, let's say, a CockroachDB, you can write down the... You can evaluate them and pit them against each other scientifically,

point by point. So when it comes to technical decisions, and most decisions in our environment are technical, there are rarely any conflicts there are debates you know space versus tabs you know there are debates but you know we lean towards that extra point one person that makes slightly more sense when it comes to for the record it's spaces just wanted to set the record straight but sorry sorry please please continue yeah i mean it was one of the biggest question marks in my life and i just gave up i don't care you know go fmt pep8 whatever i i just i'll just pick one of those tools and stick with it and the rest of us will also stick i really

don't care what space is the tabs anymore but i used to before I stopped caring, for the record, it was tabs. With sheer beauty, you just can't. I'm shaking my head. After we are done recording, I would love to watch some of this. Minus two respect points. Sorry, please continue. But yeah, so pragmatic decision there. Spaces tabs don't work as long as there is consistency with everyone's work.

And we have some sort of central formatting tool that does it it's not a headache so you know we are over that and that that's really how that's really how objective decisions are made now there are ui ux usability decisions that may not be so objective there we again present our cases very subjectively. This is easier on the eyes or this looks better and the debate could go on for hours, minutes, hours, even days but in the end there's always someone who has a point extra 0.0 on a 0.1 percent. The rest are respect that enough to be like ah yeah whatever so be it so yeah so there are no conflicts there are debates that's pretty neat uh a very well-oiled machine well humans is what i should say uh now many of these things like you mentioned when it comes to

providing that feedback evaluating people and how well they are performing, being able to set certain principles for the team and setting those expectations and the foundation for this is how we function, this is what you're expected to do. A lot of these things typically are things that managers do and considering that when you started at Zerodha, like you said, you were a hobbyist programmer. So how have you grown as a manager, even though your role is... I mean, I've seen your GitHub commits, and I don't know how you do it all.

But there's a whole responsibility of doing things outside of technology part two. What factors have contributed to your growth in that area so i i was a lone hobbyist freelance developer before zerola and that's what i wanted to be and zerola only really happened because you know the human element checked all the right boxes i didn't want to have a job i didn't want to be in a corporate environment with hierarchies and structures. And with Nitin, I figured we could just build a team, just being humans. And that's exactly what we've done in the tech team

and outside the tech team that, you know, Nitin's team also. It's exactly that. So, it just has, it's been, it's really based on, I know common decency, have mutual respect, sorry, have respect, be decent, be objective when it comes to technical decisions, be as unemotional as possible. And you just deal with people decently. I never managed a team. I had no formal experience in managing a team whatsoever. So when we went from one engineer in 2013 to two and three and four,

it was very organic. When there were the two of us you know again behave with each other decently do stuff technically and that extended to three people and today that has extended to 33 people now the tech team is 33 people two designers two mobile developers and the rest of us are you know developers and and one person who helps liaise with other departments. But the rest of the company has grown quite a bit. There's, I think, 1,100 people outside the tech team. Compliance, legal, accounts, stuff like that.

So many of these philosophies that we've developed in the tech team and the you know the core team that nathan works works with we try to extend that to the broader teams also but it is so difficult it's in in the other departments there are hierarchies there are levels there are managers you know in the support department there are managers it is impossible you can't it and a lot of that is really mundane work right answering financial support calls is mundane work how could there's no creative spark there so this whole thing about building a nice human team of people who like each other work well

with each other who have mutual respect. I think is not universally scalable. That's a realization that I've had. When there is a shared goal of being creative or having fun or solving intellectual challenges, etc. It's possible to find people with the right wavelength who can stick together. And when you find people with the right wavelength who can stick together. And when you find people with the right wavelength, it's possible for the high probability that all of them can behave decently with each other. So, I mean, I don't really like, I don't think I'm a people manager.

I just know how to work well with my team that has evolved with these very specific philosophies in place. If you put me in a different environment and said, now manage this team, I might just fail miserably. I have no idea. But I mean, I don't have that much experience, but I do definitely remember the two different archetypes of a great manager versus a great, like a technical leader, right? Because I do feel that while the manager does make me feel like the

team is well run, like the leader, like the technical leader is like the type that really makes me feel like I want to like be like them. And I feel like being in that position where, right, like being an IC as well, you're really, and also you're like kind of leading by example, like literally leading by example by doing like what's being done, right? So I think that's very powerful. But to your point, it doesn't scale. But I guess technology, we are in a lucky position where like if you're doing things

right and you're investing all the time. Anyways, those were all nice things. Rodney, you should ask some hard questions to uh you know balance that out um i was gonna do a hard pivot and ask a couple questions which is uh some things that some of our friends have uh poked us to ask it's like hey please make sure you ask this i'm like we'll try uh one of the things so i wanted to read one aspect which was mentioned on your uh blog and this is something that nathan also wrote i think in december uh 2021's blog where he mentioned that uh brokering is going mainstream or something to that effect uh he mentioned that

it took zerodha almost 10 years to get to 2 million customers and around the time when covid hit uh you added 6 million customers in a span of 18 months and you'd also mentioned in one of your blogs that you're doing around in January 2020 you were doing around two plus million trades and in three months you're hitting what six plus or seven plus million trades now that's a step function growth or a hockey stick growth, whatever you want to call it. You've written in great detail about the way you build systems that allows you to scale with common sense. But when this happened in that span of time, especially as your team is also going through this change of having to work from home, and all the other systems allowed you to do that

in a very good way way what parts of your stack had to evolve or change the most to meet with the scale even though the rest of the stack was working really well working remotely not really a technical aspect but from being a tight knit team who sat together had fun together

worked together uh going going full remote overnight that was a bummer that significantly affected how we communicate communicate i mean we kind of have the hang of it now but i would still think some 50 60 percent of all the time we spend communicating with each other on over calls and on chat is just really wasted overhead. Decisions could just be made on the spot and we would debate right there and be done in five minutes. But now it takes hours or even days to organize. And that element of spontaneity, which is critical to innovating and big breakthroughs is I mean it's it's lost most of it is just lost if you're not sitting together you can't really have spontaneous conversation so I feel that has definitely affected that's the part of our human stack

that has been significant significantly impeded the stack, thankfully the slow, gradual, first principles based work that we did over seven, eight years put us in a position where when we went from two million users to eight million users, we didn't really have to do... there was nothing catastrophic that happened. Of course we had to tune a few knobs but we didn't have to change things overnight. We didn't have to do those things. Now, there's no formula. It's just that all the common sense decisions happen to pay off. So we scale well, but the biggest bottleneck in this industry is the legacy dependencies. You depend on multiple exchanges. You depend on

lease line providers to connect you to those exchanges each new lease line has super limited capacity like thousand messages a second that's it if you want another if you want to send thousand more messages you have to get another lease line and each lease line takes three or four months if you're adding more and more users you need to increase capacity I mean you're adding more and more users, you need to increase capacity. I mean, you're just, there is absolutely nothing you can do, right? So this kind of bit us, but thankfully we've been, you know, working on setting up parallel infrastructure for lease line connectivity over like one, two years anyway.

It took two years to get, you know, four lease lines connected at a new rack in a new DC. But that kind of coincided with this explosion in 2020 March. And that was a highly experimental setup. This is the core auto management system piece that connects all these platforms to the stock exchanges. And we flipped a switch on a Saturday, went live on this highly experimental only beta tested setup and went live on a Monday with, you know with insane traffic. And we were all crossing all the fingers that we had.

But that was the closest we got to an implosion. But that worked out. So apart from this, where it's really difficult to scale the physical constraints horizontally. Everything else was manageable. We didn't go on a hiring spree. Only two people joined our team after that, after the whole 2020 growth. And they joined for different reasons. So I think that is an indicator that none of the systems really needed any sort of massive midnight oil burning or overhaul.

And we just continued what we were doing anyway. Nice. Since you mentioned one thing about the leasing lens, it reminded me of, so we asked this question to every guest, and I think this feels like the right time to ask that. Can you share what's your favorite misadventure been? And this could be during your time at Zerodha before tech-related, non-tech, doesn't matter, no rules here. But what would you say is your favorite misadventure? I think that's a bit oxymoronic.

I don't think these misadventures, especially in this industry, can be your favorites. They're always horror stories and you really don't want to think about them. And we've had many close calls. I don't know if this classifies as a misadventure. While all of this

was happening in March, April, May 2020, our instant account openings, the whole process is online, finishes in five minutes. There's MakerChecker, KYZ, you know, really complex flow, which we'd optimized. So accounts that would get open in a few hours started piling up for like 15 days because, you know, the new users signing up. It just exploded. And it wasn't exactly a technical issue, it was a human process issue. This industry wasn't really set up to handle so many online signups.

So while we were grappling with all of that, trying to figure out resources, better KYC processes, etc, the markets, regulator and our scale, the number of orders were picking up like crazy every day. And we were monitoring systems, monitoring them 24-7 just to be sure, but they were scaling all right. The stock market regulator came out and said, starting next month, I think the red line was three weeks or whatever, how you sell stocks on a trading platform completely changes forever. So how it has worked over the last 30 years, you're going to change that in the next 30 days.

So you log in, you click buy, it's bought. 20 milliseconds is bought. You click sell, it's sold. 20 milliseconds. Now each sell had to be authorised on an external payment gateway sort of a flow. A pop up should open, users should go authorised. They'll get a message SMS validation message they had to enter that they'd have to enter that OTP on the gateway authorize come back to the trading platform and sell that is just

ridiculous right that changes how everybody has sold stocks in India until that point forever that completely changes the architecture of all the you know front facing trading platforms all the architecture of all the front-facing trading platforms, all the UX product flows, the entire backend. You can no longer send sell orders to the exchange directly. You have to integrate with this external thing. And if there are a million people trying to sell stuff at the same time and this gateway which is hosted by an external depository, if they can't scale and send out a million SMS, suddenly the entire market is choked. There were so many question marks and three weeks to build this and go live.

And we wrote to the regulators, you know, sent deeply technical points. But that's the thing about Indian regulators. Once they decide, they decide. But that was quite insane you know i uh after art systems had become stable maybe by around 2019 i'd stopped working 14 15 hours a day and i cut it down to like 10 hours but this one piece so this one piece no like 15 16 hours a day working on it every single day and that was just me working on one piece. There were other people working on other pieces and everything had to come together on the 21st or 30th day and we had to go live on a morning.

And that was a proper misadventure on so many levels. So many things broke. So many expectations failed. And we're talking about integrating with a legacy institution who's still using triple encryption. Encrypt, put that encrypted payload in a JSON structure, convert that to B64, encrypt again. They call it triple encryption.

With some cipher that was obsolete 22 years ago, which none of the popular libraries support. So you have to manually pad the ciphers by reverse engineering. And you have 21 days to change your entire business and trading platform and the backend systems and support systems and educate your users saying starting tomorrow when you sell stocks, they won't be sold. You have to do this thing. Oh my God, that was a series of misadventures.

And that is just one of the things about this industry here. The regulatory volatility and regulatory uncertainty is absolutely insane. All your grand future plans that you've built today to grow your business might just be killed 24 hours from now when the regulator comes out with a circular. That's how it happens. You get a PDF and it says change everything so yeah series of misadventures that i don't really know how we scrambled but we scrambled and everything fell into place on the deadline day in the morning markets opened and nothing crashed in

work people were of course and i'd like i can't sell my stocks but we had to slowly educate them over weeks and months. Oh, wow. Yeah. The most recent one. Certainly sounds very horrifying. Very risky. Oh, wow. Thinking, or since you mentioned regulation, I heard this news recently, and actually someone even, some of our friends mentioned this as well,

that India came up with the financial budget and they mentioned crypto and I know I think Nitin mentioned on social media that when it's legal and the regulators allow it Zerodha will be ready with it how are you thinking about that from a tech perspective not thinking right now at all so if if if it becomes regulated or legalized firstly I think that is that completely defies the point of decentralized if you know if regulators say these are the three four cryptocurrencies we're going to list them on the regulator stock exchange and you can leave you can only

trade there that is just like any other future options commodity. It has no crypto, it has no distribution or decentralization. But if that ever happens, I'm guessing, it's unlikely to happen, I think. But if that happens, it'll just get listed on the existing regulated stock exchanges just like all other stocks and they should implicitly start showing up on trading platforms anyway. So, yeah, that's my assumption. I could be completely wrong.

Yeah, and the budget did not specifically mention crypto. I thought it was quite interesting how they framed it. They said digital goods. Whether they be backed by cryptographic methods or not, profits from transaction of digital goods incur a 30% tax. That's quite clever if you think about it. Now, it's a little dubious what exactly is a digital asset. I mean, if you, yeah, digital asset.

But yeah, they very cleverly avoided cryptocurrencies. I was going to add a dumb joke about how both of us don't know anything about crypto. We're just asking you for SEO, but we can skip that. We can skip that. I only have a very surface level, naive understanding of crypto. Yeah. So I don't really know anything about crypto either. Well, that makes me feel a little bit better about myself.

Well, we're coming to a close. And Kailash, you have been extremely generous with your time. Before we close, though, I have one question which I would like to sneak in, which is seeing how you have developed a tech team, seeing how you have been doing engineering it's apparent that you have thought about a lot of things and you think deeply about systems people how you want to work together have there been any people who have influenced your life the most or influenced you the most or your opinions or doesn't have to be people a company

a resource someone who has been an influence on you i think i've picked up bits and pieces from many many many different people so one of the things that i really like uh i like reading history completely random history like the memoirs of a certain person that we don't really come across. And I find it to be very revealing. These memoirs are also, of course, biased also, but good accounts of history. And it really demystifies the lives of many people that we look upon. At the end of the day everybody i mean everybody is a human being they

have the exact same human feelings love hate jealousy you know insecurity so i think once it real that this is one of those things that really led to me my being an absurdist you know full-blown because once you realize that doesn't matter if it's a president or a prime minister or astronaut or the biggest business person ever, they're just like us. They're super insecure, sad, jealous, you know, angry, irrational. So once you accept and realize that,

it then clicks. So I've picked up lots of things from many different people. But there's this one book that I think completely flipped me over into my absurdist worldview. It was Phantoms in the Brain. Beautiful book.

I've read it like three, four times over the years. I think I first read it in 2008. And my professor gave it to me saying, you might like this. It's about brains, intelligence, you know, consciousness, neurocompetition. That's what that was our research area. And V.S. Ramachandran's book, he's a neurosurgeon. And that was really the tipping point for me that, you know, everything is just in here and all the our entire worldview is just a figment of our imagination so all the reservations that we hold all the deep convictions and crazy strong ideologies that we

hold it's all here i could just uh once this podcast ends i could be getting up i could you know tip over bang my head and i'd be a completely different person it's just just that simple right or the president the most powerful person could just you know slip in their bathroom and you know be bedridden changing the course of history forever so if you read history you'll see that many humongous you know historical events that have shaped the world have just been really random coincidences so this book just really drove the fickleness of our own minds and our own preconceptions right into me. And it sounds, it sounds cliche and a little hand wavy, but you know, that sense of being somebody, the sense of the somebody-ness in me, it just vanished.

I'm like, shit a I'm a nobody and when I snap back to it saying oh I'm a program I do these things I have my responsibilities but I snap back again saying at the end of the day I'm a nobody I'm here today I'm not here tomorrow and anything could happen to me at any moment you know I could just have an aneurysm and all my life goals will be wiped out. So what is even the point of having long term life goals? Just do decent things, be decent. That's really the absurdist worldview. No long term goals or really strong ambitions.

Just do decent things. That one book was really the tipping point. And after that, I've read many similar books, just reinforcing the idea that it's all just... Yeah. Thanks for sharing that. We'll certainly link it in the show notes for our listeners to check out. And I think that's a great place to close. And Kalash, once again, thank you so much for taking the time. This has been awesome. We've learned so much just in this conversation and I'm sure so will our listeners.

Thank you so much. Thank you so much. Thank you so much for having me. Thank you. Hey, thank you so much for listening to the show. You can subscribe wherever you get your podcasts and learn more about us at softwaremisadventures.com you can also write to us at hello at softwaremisadventures.com

we would love to hear from you until next time take care