Yash Nelapati was the first employee and engineer at Pinterest.
We spoke about what he was doing before Pinterest, how Ben Silbermann got him to join, and how he built Pinterest’s first version and helped scale it to hundreds of millions of users.
Yash is currently working on his own startup, MakersPlace.
I actually didn’t get a job out of school. Probably one of the few who didn’t. I hadn’t done many internships or anything. I was doing a bunch of freelance work via Craigslist and ended up meeting a design firm called Design Reactor.
At Design Reactor, I was hacking a bunch of PHP scripts together to make a CMS for them. A lot of our work was focused on enabling virtual conferences (think what Zoom does now).
The job wasn’t really challenging so I was spending a lot of my freetime on open source, writing Django applications.
I would go to a lot of meetups, work on problems like open polling, and also spend time studying real time chat applications (Tornado had just been acquired by Facebook at the time).
It was at one of these meetups where I met the Pinterest folks.
It definitely wasn’t an easy decision to join them.
Ben’s pitch to me was that people love to curate and collect. And he wanted to bring that into the real world (via Pinterest). After I met them at the meetup (Ben & Paul), we went for a coffee and I visited their office in Palo Alto.
It was a pretty bad, shady office. They didn’t have a restroom. So my first impression of them wasn’t great.
On top of that, they were using a CMS called Magnolia at the time and one day they lost their database. And their backups were all corrupt for more than a year and they didn’t even realize. That itself could have shut them down because of very poor engineering-dev ops.
But Ben was really excited about the idea. I also knew that I would personally use the product.
Again, though, I wasn’t ready to work in that office and I also wasn’t too sure about the team. They actually pivoted into Pinterest. There were around 7 iterations, and the 6th iteration was centered around brands coming to pin stuff and users then adding some of items to their collections. But that didn’t work out because it was a classic chicken & egg problem - brands wouldn’t come if there were no users and vice versa.
Anyways - when I met Ben, I didn’t respond for a long time. Because I wasn’t sure if I should take this huge risk and as I mentioned, they had no money. They had been trying to fundraise but that wasn’t going too successfully.
Another doubt was that I was on an H1B visa and pursuing this opportunity would mean putting that at great risk too.
I basically didn’t respond to Ben for 60 days. But he kept calling me 4-5 times every week. He was really persistent. And I didn’t like my current job. I thought that even if they didn’t pay much and if the idea didn’t work out, I would enjoy this opportunity. Because I would be the one building the whole platform from the ground up.
So I ended up saying yes.
Yeah, I mean I don’t have the best of relationships with my parents, but I told them I’m doing this crazy thing and I’m just going for it. Almost every week, they’d volunteer to send me money, and I’d always refuse. That was the last thing I wanted.
I was basically in a do or die situation. Because if I went back home, I’d probably have to work with my dad (Yash’s father is an entrepreneur) and I didn’t want to do that. I had decided that I wanted to forge my own path here in the USA.
At the same, though, I didn’t view taking the Pinterest job as too “risky”. That’s because I’m an only kid, I didn’t have anyone to support, and I don’t need much to survive. So the bare minimum was enough for me. And I’ve always been like this - even in school, when there were internships available, I didn’t really take them.
I was scared of getting stuck in a particular path. So if I did an internship in testing or some job which is tailored to IBM (or whatever company), I was scared of just having to pursue that field for the rest of my career. That’s what happened to a lot of my friends. So their next job is always going to be based on what they did before (hey, you have experience in X, now go do more of X).
So instead I did a lot of open source. Read a lot of google code, even though I didn’t understand most of it. I coded a lot of IRC channels and used to ask lots of questions. People would get annoyed (they’d say things like: “Dude, you’ve @ mentioned me like 5 times now - maybe you just won’t get this concept!”).
Oh, and right after I joined Pinterest, we decided to move into Ben’s apartment because the office was too expensive, so I used to tell my dad that I was working out of this random dude’s bedroom. And he’d always complain and ask me why I couldn’t just be like my cousins and get a job at IBM or Google. But yeah, I was pretty set on following my own path.
I was very happy. There was so much technology coming out at the time - for example, almost every week there’d be some new type of database on Hacker News. And I would try to use everything - so even though I write to the original MySQL, I’d try to replicate it in Redis or Cassandra, for example. But almost everything would break. And that was one of my big lessons at the time: if it doesn’t have years of experience or years of people’s time making it better, don’t use it because your database is your bread and butter.
I’d start coding at around 9am and then we’d deploy to production at around 5/6pm. The site was small enough at the time that I could whip out features really fast. We’d get emails from users (mostly Ben’s friends) about bugs on the site and I’d fix all of them. As time went on our users became more forgiving as well.
Ben was really stressed trying to fundraise, but together we’d enjoy fixing bugs and watching the site grow. I think one thing that Ben learned as a leader was that it doesn’t matter what level you’re at - as long as you’re able to take the right feedback and learn from it, then you’ll get to where you need to be eventually.
Overall, though, this was a tough phase for us. The first 1000 users could all be traced back to someone we knew - we were still in a very small bubble.
There was this lady, Victoria, who ran a blog called SFGirlByBay, She ran a contest where she asked for people to link their submissions on a Pinterest board. And so she ran this invite on her blog and it brought around 1000 users to the site.
Those users really liked Pinterest. They started creating their own personal boards and kept coming back to the site. That’s when we knew we were onto something. To be fair, though, Ben always knew that the vision of the site would appeal to people. He had always collected stuff as a kid. We were just unsure on how to execute that vision in the best manner.
I hope you're enjoying the interview. I'm releasing an interview with Flexport Employee #1 next week. Sign up below to get it delivered straight to your inbox.
So this was around March 2011. These new users loved Pinterest (they were pinning a lot) and the site would often go down because of traffic. It was really exciting.
It was definitely more organized. Ben was the gatekeeper for the product. Everyone wanted to ship a lot of features at this point, but I had to say no. There were about a billion rows in our database so we had to shard it. I was barely keeping the site up whilst I was rewriting the entire sharded infrastructure.
Most of the new folks we hired were writing unit tests because we didn’t have unit test coverage at the time. We also hired dev ops folks to help with this new infrastructure I outlined.
The company also didn’t have any engineering leadership at the time because none of the founders were engineers and I was too young to take that role.
Now that I think about it, 2012 was a nightmare year because I was on call 24/7 and was not sleeping enough. We also used to put the new engineer on call from 2am to 6am and then he would restart the servers whenever anything would crash.
Yeah, I was still loving it. I would say that every engineer should try to experience working at an early stage startup at least once in their career. I wouldn’t recommend doing it more than once though!
Back then, I was just a really excited young guy trying to learn as much I could. I was fine learning things on the fly and always looking to use new technologies. Now, as a more experienced engineer, I’m much more measured. I prefer to predict things before jumping right into the fire.
It was tricky because even when I was hiring engineers and interviewing them, we were still in growth mode and the site was going down a lot. So for 6-8 months it was a lot of push and revert.
But after that period, I started to enjoy the hiring part a lot more. I started going to schools, hiring students, and then watching them become full-time employees. It was satisfying to watch them grow and additionally I also learned a lot about myself as an engineer and mentor.
One of the biggest learnings for me personally was discovering whether I should stay as an individual contributor or become a manager. Initially, I wasn’t very good at communication (now I’m much better because I’ve had the opportunity to learn) so I always shied away from management. And once you shy away from management, your role in the company becomes really fuzzy - for example, you stop being invited to certain important meetings just because you’re not a manager.
In general, I see a lot of smart people who don’t want to go into management and stay as individual contributors. That was definitely the case for me: I thought that I was good at managing servers, but not people.
A lot of these big curators on Pinterest, who have millions of followers, whenever they discover a cool blog or a recipe, they pin it on their board. And all their followers see this on their feed and can then go to the original source and view the content.
But when these curators try to differentiate themselves and create their own brand, they start to question why they’re giving away all their traffic to an external piece of content. So instead, they just point the link back to their own blog or site and instead review whatever it was they had originally linked.
So what you see here is that the original creator gets nothing. Attribution is really hard and the internet wasn’t built for attribution. Content gets stolen, screenshotted, and lost all the time.
When I left Pinterest, I knew that I wanted to tackle this problem of content and attribution. Blockchain had also been growing at the time and I was interested in it.
So that’s how the idea for MakersPlace was born. We’ve started with art because it has some really valuable properties like supply and demand, ownership, royalties etc.
If you enjoyed reading this, I'm releasing my interview with Employee #1 at Flexport within the next week. Subscribe to my newsletter to get it delivered directly to your inbox.
Some people call it a newsletter - I call it a good time. I write about tech careers and how you can get ahead in yours. It’s my best content (like this case study) delivered to you once a week.