Nikita Shamgunov
Appearances
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Then Google built a competitive product less than a year after we started a company that's called AlloyDB. So Google was turned off. Cloudflare never happened, but then Vercel happened. And so we partnered with Vercel and we launched a product called Vercel Postgres. And then the third part of company building was basically spinning up an enterprise sales force.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
We're past 5 million in ARR, but we are not starting an enterprise sales force. So that's a deviation of our original plan. And of course, the reason that there is a deviation from the original plan is the change in the market and all the activity around GenAI.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
This is my second company. And in the first company, we focused on competitive programming. And so we hired a bunch of people from kind of world champions in International Olympiad in Informatics. And there's another programming contest. It's called ICPC, which is a student contest. And I think that strategy works. And the reason it works is you have an asymmetric advantage for talent.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And it doesn't have to be competitive programming, but it has to be something that you have advantage over the rest of the world. Keith Rabois said that startups are built on undiscovered talent. And undiscovered talent is either young, is it somebody who hasn't been discovered by the world? And so in the Silicon Valley,
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Traditionally, undiscovered talent is new grads or people very early in their careers. It also could be people with non-traditional background. So for example, the leader of the very popular Gen-AI product is not coming from like a top school. And so sometimes you get those kind of craftsmen that grew up in a different environment, but they're extremely talented.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
But the world hasn't learned their name just yet. At Nian, our advantage is the fact that we are fully distributed. And it went out of vogue in the venture community. And there's a lot of like back to the office type energy that is happening. But because we're a Postgres centric company, we need Postgres committers and contributors. It's fairly natural for us to be remote first.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And we built a critical mass of systems engineers in Europe. So that allowed us to be a really cool brand for people in, I don't know, in the UK, in Germany, in the Netherlands. And that got us asymmetric access to talent. Stas is Russian. The first four people on the team were the ones that Stas hired.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Obviously, Russia started war on Ukraine, and it became, for a US company, we could not afford to have anybody in Russia. Everybody on that team moved out of Russia, every single person, in the next few weeks after the war had started. So that kind of removed that dependency. But because Stas was a star and he was able to attract some really high quality talent early on also helped us.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
As we go along, there are a couple of things that you can also do. And you can only do it if your company is doing very well. So you either take an approach of hiring slow and firing slow, or you can take an approach of hire fast and fire fast, but you don't have much tolerance for that. And I was reading that book that's called Amped Up by Slootman.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And he was saying that contrasting data domain and ServiceNow, and in ServiceNow, he was a lot more aggressive in terms of both hiring fast and firing fast. I have some kind of similar approaches. At SingleStore, we hired extremely slow. We had very involved interviews. And it means we're hiring much faster when we used to hire than a single store.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
But we also have much less tolerance for underperformance. And I think the net result is roughly similar in the quality of talent. The DNA is different because there was a lot more systems talent at single store because it was building software and on-prem software. And here we're running a service. So there's a lot more kind of like product engineers and SRE DevOps type DNA.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Because it's remote, the talent for the most part, more senior, because if you build in the Silicon Valley, you're forced to hire more junior people and have them grow really fast. And they mostly grow by osmosis. And the reason to that is like, it's hard to pull the top tier talent out of good places because good places make sure that top tier talent is very happy.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
We hire and fire faster at Nian versus single store. And again, I'm very happy at the quality of the talent that we have at Nian. I'm not a huge fan of remote first. I'd rather prefer everybody in the office. But we're also driving a bunch of benefits by being remote.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
The original idea was to build an open source alternative to AWS Aurora. The key idea of Aurora, which we took to the next level, was separation of storage and compute. In a year, we were ready and we started to onboard some internal users. As we started to work on the launch, it leaked, right? Because we didn't hide the website behind a password. So suddenly posted in Hacker News.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Naturally, in a startup, you will have a handful of people that are responsible. Like these are the people you call to if there is a fire. And as the company scales and the user base scales, and now we have 5,000 databases spun up every day on the platform, you're going to get fires. That's just what's going to happen. And then you keep reaching out to the same people.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
My co-founder Stas was one of them. Eventually, you start burning those same people out. And there are ways to address this. And typically, you address it by starting to distribute the workload or... You address it by having stronger team leads and whatnot. From the CEO standpoint, you can think about it as you have a bunch of balls in the air.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
If the company is growing exponentially, the number of balls in the air starts growing exponentially as well, and then they will just start falling. In our case, the symptom of that was the stability of the service about a year ago. So 12 months ago, all hell broke loose as the number, the usage on the Neon platform was growing exponentially. And then we started to have outages one after another.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And our competitors started to make fun of us on Twitter by, like, referring to us without naming our name. And it was just, like, it was embarrassing. By that time, we were already for self-postgres. Guillermo was sending angry messages to me every time we were having instabilities. The classic from the CEO standpoint is once this is happening, you start building your staff team.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
That's when I started. I opened up a number of searches and we brought head of engineering, head of product, head of marketing. And then we were like, do we need a head of finance? Do we not need to get insight? I was like, fuck it. We're just going to bring a head of finance as well.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
The learning there is once you start building a staff team, if you have some folks on the staff team who are professional executives and some who are not, the ones that are not better step up really quickly. Or you will have weird staff meetings going on where the bar for communication and getting things done is different and uneven.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Once you flip the switch and start building a staff team, I think you're better off just having a professional staff team. And that's what we did in the beginning of this year. And now the service is super stable. With a capable staff team, you're much better positioned to solve problems in a very robust way.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
It's mostly the team, the team that puts max effort into the opportunities that the world presents to us. That's how I think. And I read this in a book that's called Wooden Leadership. It's like a very old book about basketball. But that's how I think about the team and the people I work with. Life is too short to work with people who lack passion or not talented in any way.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
The greatest joy in life and when you see people grow and do things that thought were not possible for them. That's what I'm proud of the most.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
There are a few folks that are not here on the team anymore that I wish I had not hired. I wish things went somewhat differently. Everything that we do and it worked, I wish we had started sooner. And I wish we had started the Gen AI effort sooner as well. Generally, I think that's another thing that I learned from that Woodham Leadership book. The team who makes most mistakes usually wins.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Just being more aggressive in general and allowing yourself and the team making mistakes is, I think, is a very good thing.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
and we had a lot of traffic coming to the website. Right before that, we started to see our GitHub repo, and we were building in the public. The stars started to go vertical. We started to get validation that we're onto something. My name is Nikita Shamgunov, and I'm a CEO and co-founder of Neon.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
I think we're living through a transition today of software. That transition is driven by AI, by large language models. While Nian is an incredibly useful database service, I think the future is a lot more in a platform and in the agentic space. In the future, I strongly believe, A, we're going to have a billion developers.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
So right now we have 20 million professional developers and another 80 million or so tech people, product managers, designers, sales engineers, whatnot. And we're going to have a billion developers. The other thing that I strongly believe in is that each of us will have a bunch of agents that are running around the Internet doing tasks for us.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And it's the amount of information that and like various accounts that we have on the Internet is like absurd. Right. And especially those who have kids. I do.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
there's all this like flow information coming from the school and things that that you need to get done and just like the amount of digital work is huge so we're gonna have all these agents running around internet doing work for us and all of those agents are going to be built somewhere on on a platform
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Definitely all of them are going to need the database, but they're also going to need a bunch of other things. So I think the future of Neon is more of a serverless agentic platform and not just a Postgres database. And that's why I'm really rethinking part three of the original master plan, because in the original master plan, I'd be spinning up an enterprise Salesforce right now.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
But that agentic opportunity isn't front of us. We're seeing software being built on top of the Neon platform. And so we want to catch that wave. And I think that's the most important thing right now.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
It's many people. I learned a lot from Vedat Khosla, working for four years at Khosla Ventures. And it's both a mindset, the Swiss-taking mindset, and thinking about the world in terms of asymmetric opposites.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And so taking risks and and thinking about like investments and return of investments and you putting time and energy into things and you want to put them into things that are higher, potentially higher in risk. But they have like gigantic asymmetric upside if they work. And it's OK that not all of them work. You just need a handful of them to work.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And then how you think not just one, two years out, but try to think about five, ten years out and create a thesis and then be directionally correct. And if you're directionally correct, you'll arrive into a good future. So that was a big influence. And then I read a lot of books. I read books about leadership. I read books about founders. Obviously, I read about Steve Jobs. I read about Elon.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
I read their biographies. Elon's absolutely a hero. Bill Gates is a hero. Sergey Brin and Larry Page are heroes. I'm trying to learn as much as possible by reading and learning how they operate and why they've built such incredible businesses.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
I think it's going to be tailored to the individual. And depending on the stuff that he or she is showing to me, if something's groundbreaking, it will take care of itself.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Hopefully the person who's showing me understands the potential of this thing and A, not going to squander it, B, not going to be pushed around by whoever else is around it and take full ownership and unlock the full potential of both the product and that individual. I would also try to make sure that this person understands the importance of other people believing in him or her.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
When I work with entrepreneurs, that uncompromising belief that they will be able to accomplish whatever they set out to do is actually a great source of energy. And it's like an important ingredient to the overall thing working and this person accomplishing their mission.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
While the first company that I built was mostly driven by my understanding of database technology and the desire to build a startup, Neon was engineered not just as a product, but as a company as well. So after my tour of duty at Single Store, where I was CTO first and then CEO, I joined Coastla Ventures, which is a tier one venture firm here in the Silicon Valley as a partner.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
As I walked in to work with a famous venture capitalist, Vinod Khosla, I told Vinod, hey, I have this other idea. Maybe we can incubate it here at Khosla Ventures. And the idea was to roll up all Postgres usage in the world. While building Tingle Store, I noticed that there is so much Postgres out there.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And the other property of Postgres is that a small startup would be using Postgres and a very large enterprise would be using Postgres. And then as a technology, nobody really owns Postgres. There isn't a dominant software company that owns Postgres code. There isn't a software company that owns Postgres trademark. So it's like open source in the best definition of being open source.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Since then, I started thinking, what are the key properties of a successful Postgres company? One thing I learned from the node is the team you build is the company you build. And so the second question is like, what is the right founding team to build a dominant Postgres company? And then I started engineering that founding DNA of the company.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
I started calling all the Postgres committers and contributors. And then I started to think about the architecture as well. The architecture is the architecture of separation of storage and compute and therefore you need people who understand storage very well and you need to bring them into the company.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
The company also depends on the knowledge of Postgres internals as well as the ability to plug it into the Postgres ecosystem and be accepted by the Postgres community. The two co-founders, Stas and Heike, are from the Postgres world. Both contributed to the Postgres engine. Heike is like a legend in the Postgres community and a committer for 20-some years.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And then Stas worked at the Postgres company, but also worked at the tech prime as well, at the largest company in a tech company in Russia at the time. And then the thing is, you need to, what Bezos says, you need to be stubborn on your vision, but flexible on the details. There's a bunch of things that we discovered on the way.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
The first thing that we discovered, it came mostly with one of our angels. Guillermo Rauch, who is the CEO of Vercel. And through his eyes, it became obvious how important developer experience and how important serverless is in building of this company.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
So if you're building a developer-first product, there's all those things across the production quality, the ability to deliver your technology in a serverless way, integrating into developer lifecycle. All of those things, they're obvious in hindsight.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
But it takes actually a good amount of effort to build and deliver both on the technology, on the experiences, on the production quality, and on the team as well. But then as we went along, we realized that developer experience, integration with JavaScript frameworks and the JavaScript ecosystem are a big deal, and that had a big impact in the company DNA as we went along.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And of course, that statement, the team you build is the company you build, can be applied at every stage of the company. And so this year, the team you build is the company you build was applied to the management team. So we have now a fantastic staff team, VP of engineering coming from Mongo, VP of product coming from GitHub.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
So we apply this team you build, the company you build, to each iteration. And the year 25 is going to be about GenAI, and then we're bringing some of the AI talents to the team as well.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
The original idea was to build an open source alternative to a very popular product called AWS Aurora. The key idea of Aurora, which we took to the next level, was separation of storage and compute. What you do is you build kind of one storage for everyone. And what I mean by that is it's one global multi-tenant storage system.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Compute is still good old compute of a virtual machine in which Postgres lives and then attaches to your storage. So the first version of that was built in about 12 months. And then in a year, we were ready and we started to onboard some internal users. And then we started working on the launch process. As we started to work on the launch, it leaked, right?
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Because we didn't hide the website behind a password. So suddenly posted it in Hacker News and we had a big discussion on Hacker News as well as a lot of traffic coming to the website. And then we also, right before that, we started to see our GitHub repo and we were building in the public, the stars started to go vertical. And we're like, wow, what's going on?
Code Story
S10 Bonus: Nikita Shamgunov, Neon
We started to get validation that we're onto something. And the GitHub repo was, again, public. And the description in the repo was, this is an open source alternative to AWS Aurora. And there was high velocity on the repo, right? So we were shipping a lot of code.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
The tools that we used also helped both from the standpoint of how fast we can build hardcore foundational technology and from the marketing standpoint as well to believe it or not. So our storage is written in Rust and GitHub has this viral loops that's called GitHub trending that are on a per language basis.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And so because there's so much interest in Rust in the world, we were making GitHub trending for Rust many times throughout the history of NEON. And that just gives us a boost and it's a self-reinforcing kind of loop. So now that repo shows up on the trending, some people follow Rust trending and whatnot. So that got us a good number of GitHub stars.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
I think we executed this part perfectly because early on in the zero to one of company building, especially for a long term systems project, the market fit is a big question mark. If you build in public, if you build open source, if you make it stupid simple to go and try your technology, your market fit signal is a lot more clear than otherwise.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
So giving technology, putting technology into the hands of people is just a very good idea.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
Very early on, I wrote in the original pitch deck, I wrote a 1-2-3 strategy. The first one was to build separation of storage and compute. That's what we did in about a year. Of course, at that time, it wasn't super mature. It has all sorts of bugs and limitations. But it allowed us to launch, and it allowed us to start to attract users.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
And once you start attracting users, they all give you feedback. And so now you have that flywheel going. And because we had a free tier and it's so easy to start, that flywheel, it gives you a statistical way because some things just don't work or people are constantly pushing the ceiling of what's possible.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
That's the advantage of having a popular service on the internet that people use and depend on. The second part was win hearts and minds of developers and establish and sign up channel. And I said, it doesn't matter how long this part is going to take, but we're doing a good job if we cross the 5 million in revenue run rate through this.
Code Story
S10 Bonus: Nikita Shamgunov, Neon
In the first two parts, I made a point that we are not going to have a Salesforce. I thought, because there's so much Postgres out there, I thought we should be getting to 5 million in ARR without the Salesforce, purely in self-serve. But we will have a partnerships team that will allow us to establish strategic partnerships with, at the time, I thought, Cloudflare, VMware, Microsoft. and Google.