Co-founder & CTO, Starsky Robotics
Kartik Tiwari is the co-founder and CTO at Starsky Robotics, a company dedicated to creating state-of-the art remote-controlled trucks to make the roads safer.
Watch the interview
Read the interview
Rohit Mittal: Hi, I'm Rohit and today we have our guest, Kartik. Kartik is the co-founder and CTO of Starsky Robotics. And Starsky Robotics is an autonomous trucking company.
Kartik Tiwari: They try to call it 'driverless'.
Rohit: Driverless trucking company. And you guys went through YC in the summer of 2016. So you've been around 2 and a half years or so. In that time, you've raised more than 20 million dollars in funding from leading VCs such as Shasta Ventures, Data Collective and YC itself. You are originally from India, you went to RET BHU, did a Bachelor’s in Mechanical Engineering, then came to CMU for your Masters in Robotics. And funnily enough, Starsky is your second company and your previous company was Updroid, right?
Rohit: Cool, so let's actually start there and then we'll go back in your background. So you graduated from CMU and then right off the bat, you started a company. How did that come about?
Kartik: So I didn't start off with the idea. One of the things that I will fill into my background is, before going to CMU, I actually worked at an automobile company in India, Maruti Suzuki. And I absolutely hated the fact that it was a large corporation and I didn't have visibility. So just the idea of knowing more and leading the charge of development was attractive to me. But that was still an abstract thought, there was no concrete execution associated with it. And one of my friends who I got to know during my CMU program actually came into the program with an idea. And by the middle of our session, he started just talking through it with some of his friends. Fortunately, I was one of them and we bounced on the ideas, and then I realized that we can work together. And we had similar interests in actually pursuing goals like these and also complimenting each other in certain skill sets. By the time we were in our second semester we already had decided that we were going to do this.
Rohit: And what was the idea itself and how did it transform to whatever it is? So you were on it for a couple of years at least...
Kartik: Yeah, I think we were working on it for a total of two and a half years. We started off with a problem that we saw for ourselves. We were part of a robotics group at CMU and one of the main challenges that we had was that the cutting edge work that we wanted to do was part of a 20% time slice that we would dedicate. The remaining 80% of the work was dedicated just to redo everything that was done before, like building the mechanicals, building the electronics, all the things that were already done before. And the actual software that we needed to build, or the actual mechanical joint or hardware that we wanted to build just formed the last 20% of the work, so that was a problem. I think that still is a problem and we saw that it was at the time when we were at CMU for our projects. And while we were thinking about it, we realized — maybe we should have a platform that we can build on, and initially, it was a simple idea. You can think about it as a Raspberry Pi equivalent but for robots, specifically designed for that. And then over time, we realized that that was still a really niche market and we started growing on that idea and built up a full hardware kit around it for people to get started on robotics.
Rohit: And were there any common use cases you were thinking people would be using it for? Are these Boston Dynamics robotic arms? Or something simpler, something to carry in a warehouse or something like that?
Kartik: I would say that that was the biggest mistake that we make in my last startup. We started off with a use case that was pretty close to our hearts because we faced that problem. The unfortunate thing was that it was still a niche product, it was a niche market — and we just went ahead with it. We never, I guess, confronted the idea that we have to talk to consumers or our customers — which took us in this vicious cycle of us correcting our assumptions or self-correcting ourselves, and not really evaluating what our core assumptions are — which meant that we had some use cases in mind, but those were not MVP use cases, which means that we had to be as big as Google or Apple to even [ph ‘rule’ 0:06:40] out that product for people to effectively use it.
Rohit: So initially, you didn't talk to consumers and you built a product because you thought you are the consumers and you were building what you wanted. The feedback loop from outside customers, who were going to probably pay for this, was not validated.
Kartik: It was not there, and that was a really big lesson that I learned.
Rohit: Got it. So that was the startup itself, but I'm also very curious about the dynamics that existed during that time. So what did you parents say? You may have spent a lot of money taking student loans and whatnot, to study at CMU, and then you just went ahead and did a startup and not get a job at some big tech company like other people do.
Kartik: Yeah! That was a really interesting dynamic. I got ourselves into huge debt while going to CMU, it's not cheap — spoiler alert. But I think my parents were supportive. We had a few discussions, I won't say that we are really well-to-do back in India — we are okay.
Rohit: So like a middle class family. Parents would have to take a loan or something like that to pay for education — the money is not in cash.
Kartik: Yeah. We didn't have spare cash to pay for a huge tuition, but I think my parents generally were supportive. This was, of course, a risk and they just believed. At the end of the day, they just believed. There's no trick to that. And that actually turned out — I hope, well for them as well.
Rohit: That makes sense. As I talk to other people, there is always this personal dynamic that may prevent some people from starting companies — like responsibilities. Paying off student debt is a big one, especially if you are coming from India and everyone else is going for tech jobs and they're making $150,000 and you are toiling away in a small office trying to start a company.
Kartik: I think one thing that was definitely not present was attraction to big numbers. Even during those discussions about what I should or shouldn't do and what my decision was, we never discussed the amount of salaries that I could get. That was never a thing. My reason for being in the US was to find a better job and to get to do something that I wanted to do and that I enjoyed doing. This was a step in that direction. And this is something that my parents and everyone whom I respected and whose opinions I cared about knew — that this was the reason why I came here and the reason why I left my job back in India. So that was not an aggressive discussion that we had. The discussion was more around what amount of risk this is and stuff around that.
Rohit: Did you set up a timeline for yourself? You're going to try for an x amount of time, and these are some of the milestones, and if it doesn't work out then we move?
Kartik: I did set a timeline for sure. There was a timeline for this particular program that I was in at CMU. We had a co-op requirement and I was co-oping in my startup for 7 months. And my initial timeline was to have clarity on what exactly is the future by the end of that co-op year. As all plans go, it was never executed properly. Even after the end of that co-op period, I didn't have enough visibility even for the next few months. On top of that, there was — as we all have, a forced timeline, which is visa-related. That, I think, played a bigger part in defining my timeline than anything else.
Rohit: Actually, that's a good segue to the next important part about visa. So you've graduated, you're probably in OPT, which is Optional Practical Training. How did you navigate the visa situations? Were there any challenges around that time?
Kartik: So OPT doesn't have a lot of restrictions in terms of work. So OPT was easier for me and we also had this time extension. So I pretty easily navigated the first two to two and a half years pretty easily. There were no big restrictions. Transferring into H-1B came with all its problems. The very first one was crossing my fingers and thinking of actually getting through the lottery.
Rohit: What year was that?
Kartik: This was 2015, I think.
Rohit: I think I got mine also in 2015, and I remember specifically that there was 1 in 4 chances of me getting the entry. So it must have been around that for the last few years, the number of applications.
Kartik: Thank God I didn't actually look at the chances - I just applied and then fortunately got it. It was also because I had to go back to India and get my visa stamped. And this was right before a big demo that we were going to do. And I thought that I would be able to come back, but they put me on administrative processing.
Rohit: So that's 221(g) — a pink slip, you got a pink slip.
Kartik: Yes. And I was stuck in India for, I think a month — working at night there and having regular video chats with my team here. People who were actually overseeing our demo were getting really frustrated because I was the person who was supposed to present, but I was not there. And in the middle of all of that, we decided that my co-founder — who is Mike, is going to present and we started and literally pivoted our plans in the middle of everything specifically because I was in India.
Rohit: So this was the year... Updroid time.
Kartik: This was Updroid time, yes.
Rohit: So because you got a pink slip, you were not able to present at a really crucial time. And I guess for people who are not doing startups, it also gets sometimes difficult to appreciate — how important some of these moments are. If your presentation goes well, maybe you will get the next round of funding — or if something goes bad then you are just done.
Kartik: Yeah, it's actually a pretty big deal. Even at YC. YC itself carried a big and heavy name. But demo day at YC still needs to go really well. You can talk about how good and interesting just getting into YC is. But if you don't do well at demo day, there are some difficult questions to it. Same thing goes with every other presentation, every other demo that we give — whether it's a big or small incubator.
Rohit: Makes sense. So you did Updroid, and then eventually you decided to sunset that company.
Rohit: Because you are on H1-B visa for Updroid, and again you didn't go to a company — you, again, started a company. So tell us a little bit about how you, again, decided to do another startup — How the startup came about, and then all the visa....
Kartik: Yeah, so this was end of 2015 and we were running out of cash. We had one last chance of going into CES, we sort of realized — at least I definitely realized that there is only so much I can do. This, generally, is a really emotional time when you have to tell yourself that the thing that you were working on and believed in so much might actually not be true. So it was an emotional time, and I just wanted one last confirmation where I would know that this was wrong. Not exactly wrong, but this might not be the right way to do it, or this might not be the right time to do it — I wanted that confirmation. So we had this one last shot of going and presenting at CES, and I [inaudible 0:17:08] however much money our company could afford, went there with all the four team members and finally got the confirmation at CES that yes, this is an effort in the main, at least from the product perspective. So I called it out and said “This is not going to work. This is actually a good time for us to call it off. We should start looking for other jobs”. I think everyone knew that, so it was a heart-to-heart with all the team members. and everyone went ahead and started looking for jobs. All the other ones were also US citizens, so they didn't have the visa overhead. I had, I think, maybe one or two months.
Rohit: I think it's 60 days, once a year H1-B.
Kartik: Yeah, but I had that much worth of salary left for me early on and I was just counting down my days. And I had a job offer from a company in New York, or I guess New Jersey. And I was planning to just go over there and work things out. It was not a job that I was looking forward to. And in the middle of that whole period, I just got a cold email from Stefan who is my co-founder, asking for advice on how to hire a technical co-founder. And I think it was also his was way of exploring the opportunity of me being a technical co-founder with him. His side of the story is also as fascinating as mine — without the visa complication, obviously. But we met up at a coffee place and ended up spending about 8 hours in that coffeehouse talking about our plan and vision for the future, what we think and how smart we are. And we realized that this was actually a dream worth pursuing. I obviously had a huge hurdle to cross — specifically because I just went through this whole emotional turmoil. Winding down a startup that you had worked for, that you had put so much blood and sweat into is a hard thing — and jumping into another startup was a challenge. So I wanted to make sure that the boxes were checked. I couldn't expect everything to be answered, but there were certain things — some personality traits or just the business model, stuff like that needed to be in place before I could take a call. Essentially, the experience with Updroid gave me a stepping stone to at least ask those questions, and that helped me make this decision.
Rohit: Got it. So all the mistakes that you made at Updroid, you wanted to make sure you don't make those mistakes again.
Rohit: And when you met with Stefan and talked for 8 hours, did he have an idea that you sort of helped him explore better? Or did you both come up with the idea together during that first meeting that went on for 8 hours?
Kartik: So Stefan had a really concrete idea of the problem. And I guess, in retrospect, that was the thing that sold me, because the depth at which he understood the problem was really good. And at Updroid, we did not really know what problem we were solving. It was a really good parallel because at Updroid, I was trying to build a really cool technology and that took me off guard and made me realize that technology is not what you should build — you should build a solution. And just having your gears switch in that direction and looking out for that helped me a lot. So when Stefan and I were talking, he had a really good idea of the problem. He had a good friend who was supposed to be the technical co-founder — they both flushed out the high level back of napkin solution for this and when both of us met— and this doesn't happen overnight, we figured out what exactly the approach should be.
Rohit: So the first meeting with Stefan went well. And you said that over time, you flushed out the approach. How much time did you take before you both decided that you wanted to start a company together and what solution you were going to build?
Kartik: We had a pretty quick turnaround in our decision. I had a chat with Stefan, I had a chat with the lead investor that he had at Unshackled, and asked about him. And he, in reverse, had a few of his technical friends interview me about my technical leadership and how I grade, in retrospect. But overall, that process did not take very long. I think it was maybe two weeks, if I'm not wrong. And then as soon as we said yes to each other, I took an impromptu one week vacation to Hawaii to just chill out and enjoy nature. And right after coming back, we actively started working on this.
Rohit: And one aspect to ask about again is — what did your parents say? Now that they have seen you go through one startup and you were convinced this was going to work, and then you did the same thing again.
Kartik: That's a good question. I think it didn't register too much. I think it was 'go with the flow', even then.
Rohit: So that was a positive?
Kartik: It was definitely a positive. I would say at a high level —not for this to sound mean, I just feel I'm lucky in this respect but building a company is not for everyone. And it's not because someone is smart and someone else is not smart. It is a lifestyle choice that you make. So it's something that you have to internalize — and not only you, your loved ones also have to internalize that this is a lifestyle that you are choosing. It is equivalent to living in the suburbs or the city. If you are a city person or a suburb person, there is no right or wrong answer. It is a thing that you enjoy, or it is a thing that you don't enjoy — and it is not for everyone. And it involves you, your loved ones, people who you care about.
Rohit: Yeah, definitely it is a preference. And sometimes it takes time for people to realize what they want. So you had already worked at Maruti for a few years and you realized — this is not what I want, and you did a startup. And when you decide to do it again, you know that this is what you want and this is the kind of life you're going to live, and you are okay taking risks. It may be difficult for US citizens to appreciate how much risk-taking you need to do once you graduate from school with a big student debt that your parents have to pay back home if you are not able to. And then you start a company...
Kartik: Oh actually, by the way — although we were running low on cash, on H1-B you have to draw a certain amount of salary — but whatever I used to draw, I generally had enough savings to still keep on paying the loan. So even though the risk was that I didn't know whether or not I could keep on paying the loan. But during the time off, I was actually paying off the loan. And by the time I was in Starsky, maybe about 6 months or a year out on Starsky, that was when I finally closed those books and everything.
Rohit: That's awesome and that just gives you another level of freedom. So let's go back — you talked to Stefan, you both decided this was what you were going to work on, and you started building the company. You got your first investment.
Rohit: Did you immediately apply to Y Combinator after that or did you first build a prototype? What was that timeline?
Kartik: This was something that I realized later. I did not build a prototype before we went to YC, I helped build the technology framework of how we are going to implement this. When I joined — I think this was March 2016, we had a month or so before we had to put in our application. So there was no time for me to build anything substantial, and it was mostly an idea at that point. We had a good understanding of the problem and of how we should approach solving it. There were obviously iterations, but I think we came in with good assumptions and all those assumptions have been proven right. There's always course correction, but the core assumptions have actually stood the test of time.
Rohit: So who were your partners? How has the experience after YC been? Do you have a team in India?
Rohit: So then I'll ask you about two things. One is about building a network here in the valley or at CMU, you can start from there. And how you go about building a team with your own.... whether you have a theory or thesis. After that, depending on what the answers are, I may chill down a little into them. And if not, I'll just ask you about advice for other founders. All stories are so unique, and yours is super-risky! In terms of the level or risk, I think it's the top one — more than ours! So I think that it would be fascinating for people to also see and learn how you think about risk. You have taken so many risks, so in that advice question, I will also ask about how you think about taking your risks and opportunities, cool? So now you're in YC. Who interviewed you by the way?
Kartik: So we had an interview with Justin Kan, Sam Altman and their CFO Kirsty Nathoo.
Rohit: Did you have just one interview or multiple?
Kartik: Just one. And I think Justin was the one who asked the most questions, because he had the most relevant information from his crews in [inaudible 0:31:49]
Rohit: Got it. So now Stefan and you are in YC. Things are going fine, you got done within... I think you raised your [ph 'C-Chan' 0:31:57] right after YC. What were the early proof points that you wanted to get to before going for the next round of fundraising?
Kartik: I won't say that we had anything that we wanted to prove, to ourselves at least because we had a pretty common-sense approach to the whole thing. From my side, I knew that the technology that we are trying to build is buildable — it's not light years ahead. There are less research problems. Maybe I have not personally implemented something, but I know that [ph 'sederar' 0:32:43] communication can happen. I know that we can build a motor that turns a steering wheel and all of that. So at a basic level, I knew that things can be done. The difference between having that confidence and converting that into a timeline that can fit in with your fundraising and your product development while you are also trying to hone in the business side of things, that's the tricky part. So on Stefan's end, he had already spent a lot of time just interviewing different players in the industry and figuring out what the constraints are, how to get to know and get to talk to people who can give us access to trucks. You can not just go and get a truck! So he was working on that end of things. We were mostly focusing on defining these milestones and trying to at least stay focused on what our eventual goal was, which is solving of the problem of removing the person from the vehicle.
Rohit: Got it. And then you sort of hit some of these milestones. And when did you raise series A?
Kartik: The first money in was right before Christmas 2017. Everything got closed and signed around February 2018.
Rohit: Okay, so it's been more than a year and I'm assuming you're building a bigger team.
Rohit: Outside of Maruti, you haven't worked at bigger companies in the valley. And really, this is a deeply technical problem — so it's not a social app. So there is maybe less UI design and all of these things, it is deeply technical. You have to make things work, otherwise people may die if the truck goes sideways or something like that. So how did you go about building the team, and how have you been managing as a first time CTO of a larger team?
Kartik: So building the team — I think there can be different styles to it. Generally, it follows the same trajectory that there are certain things you do that are not meant to scale. As the YC saying goes: Do things that don't scale, at first. So there are certain things you do that are not meant to scale, even in team building. And one of the things that I realized specifically for this — you need to listen to your gut feeling. It can not scale at all. You can not explain why you like one person vs. the other, but when you only have to hire one other person it works really well. Looking back, you can figure out a lot of reasons why a person worked or didn't work, but then you have to listen to your gut feeling. That also doesn't scale, you have to make mistakes in hiring. That would be one of the biggest things that you would learn and no one can give you an advice that you can just follow and do the right thing all the time. So I think that would be the first step to figure out. And then going forward, you have to formalize it as much as you can. I wouldn't say that this needs to be written down right away. The more you can formalize this, the more scalable you make building your team. But at the end of the day, there is a tradeoff between how fast you want to move and how process-oriented you need to be.
Rohit: And do you still do the final interview for every technical hire?
Kartik: Yes. We had two today.
Rohit: Oh, okay! We saw some people downstairs. And the more interesting one is: Did you have to fire anyone?
Kartik: Yes. So at Updroid — that was my first fire. I think it was a relatively simple one, there were no complications in it. Just because it was the first one, it was more nerve-wracking at that point. There were more difficult firings that I had to do at Starsky. And in general — your question about being a first time manager, that was a big challenge. I think my philosophy in this context is trying to make sure that we are really merit-based. At least I try to be really merit-based, which also means that there is an extra burden on me to figure out what that merit is. Someone just doing their job at a high level makes sense, but at some point you have to boil that down into how fast? Did they actually do a good job at their job? Do they share the same work culture values? Is that compatible with me or with the other team members? — All of that. Coming to that conclusion is really hard, and the first time I had to do that was actually really, really hard.
Rohit: That's one of the very common things that I've heard from first time managers — Firing someone. It's so emotionally challenging and it's so draining. You are still fine, but the person you are firing — it's very difficult for them. I think, as you mentioned, you have to objectively make decisions, especially at a startup. It needs to be merit-based. You need to make sure that you have evaluated all possible angles before making that decision. And it’s still very difficult. It doesn't mean it becomes easier.
Kartik: I would also say the same thing — you can think about this in a non-scalable fashion. You don't actually have to prove anything. It might be a good internal goal for you to have an objective set of reasons for why you are going to fire someone. But if you don't have all the highest [ph 'dortenenties' 0:40:31] crossed at a small scale, it is still okay because you don't have a concrete work culture defined. It is still subjective. You still go through your gut feeling, and if things are not meshing well — it is a fine reason, it is an okay reason to fire. Because like in geometry, small angles when stretched out too long will create a huge division.
Rohit: That's actually a very good point. That's why people say if things are not working out, fire as soon as possible. The impact it has on a team basically compounds. That's definitely something that the better way of putting it is that small differences become big later on. So now you have done two startups and hired a team, fired people — all these learnings. Just to touch on when you got started, where was Updroid based, by the way?
Kartik: In the Bay Area.
Rohit: Did you have to do something different to build a network? Because most people moving to the Bay Area or somewhere in the US from other countries may not have a readily available network to plug into.
Kartik: I don't think I did anything different than anyone else. I think coming to the Bay Area was essentially good. I think anyone who is remotely thinking about this should come to the area, although it is getting overcrowded and expensive, I still think it is a good bet. I don't know if this is an official term, but there is a networking infrastructure already available and built specifically to help entrepreneurs and I think you can tap into it really well if you hustle a little bit. If you are serious about building a company, you need to tap into that part of your skillset. You can't just be technical and sit in a cubicle and work on something and then suddenly someone will discover you. You have to network. There is a bit of brute force involved in it. At least back in Updroid, I tried a bunch of different methods. All the way from meet-ups to incubators. That's a pretty good jumpstart, YC is a big one. But if you get into any other incubators, you at least get that jump. There are a lot of learnings you get — how the world works and all of that. So there are well-defined methods that people can use for networking.
Rohit: Yeah. And sometimes that makes the difference. Like you getting in touch with Stefan and meeting the right co-founder makes all the difference. So getting out of your comfort zone to talk to people and meet them and learn new things is very crucial. So things are going well. I think you also announced a few months ago that you did the first driverless 7-miles stretch?
Kartik: This was a year ago now.
Rohit: So I'm sure the miles are increasing. So in your very risky journey into the startup world in the US, are there any specific pieces of advice you would like to give other immigrant and non-immigrant founders that are trying to make it in the tech world?
Kartik: One of the things that I would definitely repeat is that starting a company is not for everyone. And this is not at all coming from a smug or judgmental place. This is a literal fact of life. This is something that you need to decide — that this is a lifestyle that you want to pursue, and you need to be decisive about it. You can not hedge your bets at all. If you are comfortable with an 'all or none' situation, which a lot of people will not be, you shouldn't. It is a lifestyle and you have to stick to it, essentially. So that is a particular thing that I think everyone should just appreciate about startups and before getting into it, they should appreciate that this is the lifestyle that they are choosing. In terms of the risk level, I don't have a good answer for it. But a good thing to just know and keep an eye out for is the amount of risk you are taking — not just in one dimension. There are multiple dimensions of risk. Some of them can be personal, some can be financial, some can be the product, technology or business risks. Identifying all those dimensions and then figuring out which of those risks you have control of and which you don't. What risks can you actually affect? That is, in general, a good exercise. I won't say that I know where that sweet spot of decision boundary happens, it is different for different people. I think it’s just knowing that there are different dimensions of risk and you can affect some of them directly. For me, I can very much know that technology might be a risk, but I know how to solve it, or at least I know what needs to be done or what can not be done. The market? I don’t exactly know how it will evolve. That is a risk that I might not be that comfortable taking. Some people are comfortable taking market risks because they know they can do enough marketing, or they are better at just reading the pulse of the market. So they might have better confidence. So risk has more to do with lack of information for the future in different directions, and some people are confident in one vs. the other.
Rohit: So would say pick your battle. Basically, if you decide to do a startup, know what you are getting yourself into and then have a good sense of all the risks you are taking at various stages — whether of a startup or of life, and then be okay with those risks and don’t try to impact a risk that you can not change. For example, visa risk — you can not change it. If it’s a lottery, it’s a lottery.
Kartik: Exactly. There is a risk to it considering that it will be an all-or-none thing. How does that impact you? One other thing is, this is just an analogy — the reason I don’t have a good answer for this, and I think no one should have a good answer for this, is because someone advised me and told me this: If you do a very highly detailed analysis of what the risks are and the other risk/benefits things, you would almost certainly realize that doing a startup is wrong — it’s not beneficial and is not a good long-term investment. There is no way. If a startup is worth pursuing, you would always come out with the conclusion that you shouldn’t actually invest your time in it and that’s just the nature of things.
Rohit: Cool, that was super helpful and thank you for taking out the time to talk to us today.
Kartik: Thanks a lot for having me.
Ready to get started?
Achieve your dreams with a loan that fits your needs. Check loan options without hurting your credit score.
Talk to our team
Feel free to reach out with questions about anything from eligibility to loan offer. We are here to help.