From Amazon engineer to remote startup founder with Marin Smiljanic from Omnisearch

Victor [00:27]: Marin Smiljanic CEO of Omni Search tells his story of going from Amazon engineer to startup founder. Why he hired a CTO, even though he’s a developer and how he built a remote team. Marin welcome to the show.

Marin [00:41]: Thank you, Victor. Thanks so much for having me.

Victor [00:43]: Yeah, it’s a pleasure. I’ve been looking forward to this one, because it’s a great story. Let’s start from the beginning. How have you decided to become a developer?

Marin [00:53]: So I think that this trace is way back. Really. I think that in elementary school I had a little bit of a talent and interest in math. So it really came, came through that. And then during high school, I went to a really, really great mathematical/computer science high school here in Croatia, which is where I’m from. And I really found it interesting. We first learned a little bit of Pascal. I had really good teachers and then also there was just a community of people that really liked programming. So that’s kind of how I got into it. I got into doing programming competitions and algorithms. And then later on I sort of leveraged that into industry experience.

Victor [01:36]: And how did you then end up at Amazon?

Marin [01:39]: So that was actually a multi, there are many years in the story actually between starting out and there.

Victor [01:47]: You mean between grad school?

Marin [01:50]: No between graduating high school. But basically, after I had finished high school, I was very eager to get some industry experience. And so one of the first things that I did, and one other thing that I should mention is that my dad also owned a software company. So that’s always been an inspiration to go into the industry. So I would say that my sort of first big taste of big, serious tech companies was when I did an internship at Facebook, which was probably third year of college and then decided to try out a startup.

[02:27] Not building my own, but joining one that was a startup called Mem sequal or nowadays SingleStore. worked there for a while. And then sort of decided that I would like to try out one of the fangs and this was a time when AWS was really, really, really picking up, like tremendous growth. And so I was actually pretty interested in doing some part of AWS. So I applied to the S3 team and then got the job there.

Victor [02:56] Cool. I mean, especially working for AWS and obviously I guess Amazon generally hires good engineers, but AWS really highly technical. So this is great. It must mean you are a good engineer and I’m just pointing that out for later in our story, because that is super relevant. But now when it comes to Amazon or the other startup, did that help you with your own business? Do you have to learned anything that you leverage right now?

Marin [03:27]: It was a very different situation in these two and even the teams within Amazon were different situations. So for the startup part, that was actually pretty useful because it was still fairly small, which means that you can, if you’re an engineer, you still have exposure to a bunch of the other functions in the company. So you’ll essentially be able to talk to the VP of sales, to the marketing people, to the PMs and so on and so forth. So that was really interesting in that you could actually get a better feel for what they were doing and how they were doing their jobs.

[04:03] And you also could see fundraising maybe a little bit up close, although not entirely, that was mostly done by the founders. I think that for the Amazon part, I really had two teams, two main teams, one was S3, as I said, and the other one was Alexa.
So we were building email integration for Alexa. And these two were of course, very different altogether because S3 is one of the most complex technical, technically speaking systems in the world. It’s a huge like vastly distributed system, very high reliability, very mission critical.

[04:38] So that was phenomenal when it comes to designing systems, understanding distributed systems, understanding the key primitives of building stuff that scales. On the other hand, it didn’t move quite as fast because obviously, if it does move fast and something goes wrong, then you’ve got an outage of a third of the internet.

Marin [05:44]: So the next one Alexa was, I would say that that it was less hardcore when it came to the tech, but on the other hand, the effort was really to ship fast and to bring new products to the market. So I would say that these two balanced each other out pretty nicely, because one taught you to ship, the other taught you really hardcore technical stuff.

Victor [06:06]: I’m not sure how, how big, obviously, a team like the Alexa team would be, but was it like a startup within a company? Do they structure it that way?

Marin [06:14] It was at that point pretty, pretty big. I think even now they’re saying that it’s several thousands of people. So it wasn’t really a startup, but the teams were organized in a slightly more startupy fashion because everyone would have their own dedicated PM and so forth. So you would kind of have this philosophy of two pizza teams just working together on a self-contained part of the product.

Victor [06:41] Got it. Cool. Okay that does make sense. When and how did you then decide that you need to found your own business?

Marin [06:49] I think it’s always been a dream. I think it came from both my dad being a tech entrepreneur from reading, just a lot of Joel Spolsky and Paul Graham and these dudes. So that’s always been fairly inspiring to me, not to mention the standard mythology of Jobs, Gates, Musk, and whatnot. By the way, Musk, wasn’t even as relevant back then, but definitely like Jobs, Bezos, Gates and these guys. So I would say that that it’s been a dream for a pretty long time and there was a certain point at which I was thinking, okay, now I feel is the right time to do it just because if I wait longer, I might have a little bit less courage to do it later on. So I decided to really bite the bullet there and head out.

Victor [07:41]: Cool. How did you do that, the first thing that you kind of started?

Marin [07:45]: Yeah, so it was actually, the first thing that I ever started was a news aggregation and recommendation engine/app. So it basically had its own algorithms that would, yeah you had crawlers that would aggregate news from all the different sources and it would do that pretty fast within a couple of minutes after they were published at the source. And then it would learn from your clicks and figure out topics and sources and so forth that you were interested in and made good recommendations. So it was actually a pretty good product, but suffered a little bit from distribution problems and just overall standing out from the crowd.

Victor [08:27]: Got it. Did you do that on your own?

Marin [08:29]: Yeah, I did that on my own. Well, I mean, I was the solo founder. I had people working with me, including a COO who was sort of part-time as well as a really great, great mobile developer, but I was the solo founder.

Victor [08:44]: Did you start that like part-time next to your job or did you like actually quit as Amazon?

Marin [08:48]: No, I quit and this was actually like one of the conversations that I had with my boss at the time at Amazon. And this is a guy that I really love, he’s been incredibly influential in my career. So shout out to Michael, but we were basically talking about this when I had decided to start something on my own. And it was actually fairly recently after I got a promotion too. So it was actually kind of eyebrow raising, but I kind of thought that I would never be able to do it properly. And with the proper intensity, if I kept the job. And so we talked about it a bunch and he understood and gave his blessing.

Victor [09:31]: Awesome. But you say that really, you had to sunset that one, right?

Marin [09:36]: Yeah. That I worked on probably a bit over a year and it was a good product. It’s still operational, it’s got a really good engine. It’s got a good API, but yeah, at a certain point with these consumer apps in particular, you either go to the moon or you’re not going to be able to sustain it.

Victor [09:57]: So then you said, wait, let’s do B2B or how did that go?

Marin [10:02]: So, no, I think that at that point I was thinking about, we need to properly pivot and I probably need to get someone else to do it with me, a partner.

Victor [10:12]: How did you find someone? So you said, okay, I need a co-founder for the sake of sanity, accountability, what was your thinking because there are many reasons?

Marin [10:22]: That is actually a good question. Primarily for the bandwidth. I think there was a part of sanity. So you do feel that it’s easier to work with someone else and just sort of be focused on the same mission and kind of lift each other up. So that was the main motivation

Victor [10:44]: Was that the CTO person or someone else?

Marin [10:48]: Yes. So this is my co-founder Mati, who we’ve actually been friends for over 10 years now. And we met through programming competitions whereI was teaching like an advanced algorithms course at my old high school. And so he was one of the students. He was actually a couple of years behind me, but he was really, it was a talented generation and he was one of the top people there. So he really had great, great chops, great technical skills. And that’s why I’m like, okay, well I think you should join as CTO.

Victor [11:24]: Cool. And that’s really the interesting question I want to get into because I pointed out earlier, that you must be a really good developer. And you said, obviously, you still write some code, but ultimately you didn’t decide to get a co-founder that kind of has other skills. You decided to get a CTO and I would love to hear your thinking about this because I think this is super, super smart, but most people do it a different route.

Marin [11:55]: I think there’s a couple of reasons. One of the things is that when you’re building something that is highly technical, like what we’re building in Omni Search, it pays for the CEO to be technical. That’s just very important. You can have a business guy running it, but I think that you should be able as the CEO really to have a good hang of the tech. But I think that one of the reasons why the arrangement was as it was, is that I did see that it was important, like you said to have the business side well covered.

[12:34] So I’m like, okay, well I feel that I should do it. So even though I’m also a technical guy and I like to think a very good one. It’s important to have that combination where you are technical, but then where you also really deeply focus on the business side. So I thought, it was a really good combination overall.

Victor [12:54]: Cool. So you are the one who then said, okay I’ll leave at least the tech leadership to you. Obviously we still brainstorm stuff, like why not? We’re the founding team. But I’ll go take care of business development and I’ll just learn it.

Marin [13:11]: Yep. Exactly. Exactly.

Victor [13:13]: That’s awesome. How did that go? How was that experience for you?

Marin [13:16]: Well, it’s been interesting because nothing really prepares you for it. See the startup that I was at SingleStore, it helped a little, but not really to that extent because it’s one thing to passively observe how the other functions in the company do their job. And it’s completely different when you got to do it yourself, when you gotta create your own shot. And so that was interesting because it was really a baptism by fire. You can’t learn it other than by doing.

Victor [13:51]: Yeah 100%, but would you say that the fact that you understand your product so well technically, would you say you’re maybe even better at the business development than say a marketer that you just bring in with no deeper understanding of the tech itself?

Marin [14:12]: Yeah. I definitely think so. And also, there will be a time obviously when that can get delegated, but even when you’re talking about sales and marketing people, you do want people who have a fairly good technical understanding. I mean, at the very least at the level of like how APIs work and stuff of that sort.

Victor [14:35]: Yeah, totally. I think at this point, we should indeed clarify what your new product that you pivoted to does because it indeed is highly technical. So that might also play a part in that decision.

Marin [14:49]: Yep. So Omni Search is basically, as I like to call it, search for everything. So it’s basically a search product that enables you to find information inside any kind of file, any kind of content type. So audio and video are actually our most differentiated feature. And so what that basically allows you is to search for keywords and find exact moments inside those kinds of materials, where that occurs. And the motivation for that was obviously training videos at Amazon that was pretty annoying, searching through hours and hours finding information. And it just sucked.

[15:29] But beyond that really, we also cover a whole lot of different file types just in a single place. So from audio and video to docs, to text and to images, to slides, whatever comes to mind. So we completely make it effortless for the customer to extract the contents from these kinds of materials and just to make it searchable in a single place. And the one other thing that I would say is that we are primarily in terms of use cases, we’re primarily targeting site search use cases. And what that means is basically customers integrating us into their sites to power the search bar.

Victor [16:07]: Got it. So what you provide is not really an app or a widget that someone downloads in their machine, but this is really rather for other services providers, software companies that want to provide their users with a better search on whatever they have.

Marin [16:26] Exactly. And obviously with the tech being so flexible, there are many different directions that it can take 10 years from now, but for now really site searches are bread and butter.

Victor [16:37]: Got it. And here it of course makes a lot of sense that the primary salesperson is a tech guy, because you might be talking to a developer or to a CTO who’s looking for the right solution, I guess.

Marin [16:48]: Yep, totally.

Victor [16:50]: How did that pivot go? Did it take you long to figure out, because I see the similarity, news aggregation to search across sites and items. It is maybe not the same, but a similar mechanism. How did you validate that?

Marin [17:08]: So basically one of the main things was just building it out and really building an MVP, launching it, recording a really good demo and just showcasing it to people. So the demo has always, I even kind of would want to show you one if we had the, the capacity to do that, because the demo has always been the most impressive part because it really does look magic when you type something in and it just jump to the right spots. So the demo served as the lead magnet, so to speak.

Victor [17:45]: Okay. That makes sense. How did you generally, maybe, even from the first product, how did you fund that? Did you have funding or was it bootstrapped?

Marin [17:56]: So the news product was basically bootstrapped. This one was bootstrapped for a while, until August, I believe last year when we raised our preseed funding.

Victor [18:07]: Well, congratulations on that one. And from that obviously with funding, with, or without it’s you, it’s your co-founder you are based in?

Marin [18:18]: So I’m basically based in Vancouver. I’m like I said, originally creation, I’ve been in Canada for over five years now, but now I’m kind of then in between Canada and Croatia and my co-founder is here in Zagreb.

Victor [18:32]: Cool. So you’re in Zagreb right now. Awesome. Zagreb Croatia. And obviously what we do a lot is build teams in Eastern Europe, but for you, this came even more natural, right? You’re the broader term of Eastern Europe, of course. But for you, this came even more natural since you’re from Croatia. So how did you go about and you’re a co-founder is from Croatia. So is that kind of the natural direction where we’re building a team entirely in Zagreb or do you also build it somewhere else?

Marin [19:05]: I think the idea is basically to build the development team in Zagreb and then build the business side of the company somewhere in the US or Canada. I think that’s a good division because at the end of the day, like when you have sales people that need to be on calls, it’s worth being in the same time zone as the US really. And not to mention that I generally do like US sales people and business people a lot. I do think you have a lot more people that have been there, done that. You have people that have been in unicorns, that have been able to scale certain functions within a company.

[19:45] And so that’s really what motivates me on that side, but I’ve always been a believer that there’s a lot of development talent in the world and just focusing on somewhere in America doesn’t really make sense because you have phenomenal people around the world. At some point, some years ago, somebody say on Quora that 90% of the best developers are in Silicon Valley. And I thought that, thinking about it, I think it’s total BS.

Victor [20:23]: Okay. That makes a lot of sense that division actually. So are you structuring it as a remote first company or more like two hubs to offices?

Marin [20:32]: I think we’re pretty flexible on remote because a lot of people are really used to it as a perk at this stage. So we’ll have to see how it goes, because if we see that people are so much used to working from home, then we’ll have a bit of flexibility there. I do feel like we should get an office and then have at least the possibility of people working together, because I do feel that at least for the dev team it’s important, to just have white board sessions, brainstorm a little and just build chemistry.

Victor [21:07]: Totally. And then I guess that the team retreats either way, right? Probably to get everyone together in a room, but one important thing also, I mean you said that you live in Vancouver in Canada and that’s obviously on the west coast and it’s much easier on the east coast, but west coast really to Croatia that’s a nine hour time zone difference. Does that create any trouble for you or are you navigating around that?

Marin [21:35]: I mean, it’s not trivial, definitely because you need to find some overlap and that’s not going to be. Someone’s going to be outside of their regular working hours, but at the same time, I do feel that if the trust is good enough, and if the team members are responsible and I would say have enough individual initiative, then it’s possible definitely to coordinate via slack and just have check-ins on Google meet and stuff of that sort. So I would say, yeah, if you know each other and you function well as a team, then it’s not a big deal. Like for us, it’s not a big deal.

Victor [22:11]: Yeah. Obviously, because development in a big way, if it’s laid out well, at least for the given sprint, then it’s a self driver in essence. But with a bit of overlap. So we also see that work really well if it’s set up well, but then again, sometimes it of course gives you a few challenges. However, what’s your plan for the near future? What’s the long term mission and what’s going on right now within Omni Search.

Marin [22:42]: So I think that tactically speaking, we’re hiring, we need to hire a developer that’s the plan. We will also be potentially adding some more sales people to help out with the whole sales process and lead gen and all that. But we’re not I would say at this point, super aggressive when it comes to hiring, because we like to keep the runway reasonable and not be in a position where you need to raise within three months or you’re dead.

[23:14] So we’re kind of being careful and deliberate about it, but at the same time, yeah now is the time to hire a really good developer and to start working on some other features. Now, in terms of the roadmap really, I would say that there’s a couple of different things that we will be working on, just covering a lot of different data sources and making it super easy for people to ingest data into Omni Search, like for instance, connecting your drive and then having on our side, like a mini crawler that pulls everything and just ingest it into Omni.

[23:50] So apart from that, we also have certain features that are required for, so to speak enterprise readiness, permissions, and security features and that sort of thing. And then just a bunch of other standard architectural, scalability changes and so forth. Apart from this, we also want to make it super, super easy for people to integrate it into their sites, wherever their site may be.

Victor [24:15]: That’s a great roadmap. I wish you all the best with that. Sounds really cool. Love what you’re building. And thank you so much for being on the show. Where can people find more about you and Omni Search?

Marin [24:29]: So I think that the best way is to yeah, take a look at our site, Omnisearch.ai, and then also LinkedIn. So we have a company LinkedIn or connect with me on LinkedIn and I’m also on Twitter. So I think that’s the best way.

Victor [24:47]: Awesome. Perfect. Thank you so much.

Marin [24:49]: Yeah, totally. Thank you.