I received a grant to build NexuSelf. At that point I already had an app. Not NexuSelf as it is now but a general fitness app, for everybody. I built it myself, coding every day, often pulling all-nighters, 12 to 14 hours at once. I was doing everything, talking to users, getting feedback, shipping. I was pretty good at it, actually.
Then the grant came, and I thought: if I can hire someone professional - someone with an actual technical degree, years of experience - we could move twice as fast.
Looking back, I think I underestimated myself. I assumed a professional would automatically be better than me because they had more formal experience. That was my first mistake and a lot of the other mistakes started from there.
I Hired Before I Knew What I Was Asking For
I started talking to my friends, searching for software engineers, and machine learning engineers. I found a few through freelancing platforms. I hired one engineer to work on a food-recognition system, and another to work alongside me on the core product.
I didn't really understand at that time what engineers like this normally cost. I had no idea what an engineer with let's say eight years of experience would cost. The first one I talked to quoted around ₹6-7 lakh for the work. I thought that was way too much, tho there were other engineers who were willing to work on less amount but I wasn't sure about them so we went back and forth, and it landed around ₹3.5-4 lakh. The AI engineer for the food-recognition cost about $500.
The first engineer I brought on for the core product was just too slow next to my own pace. I paid him around $300 and told him it wasn't working. Then I hired someone else this time actually good, but he was expensive and also slow. We spent a lot of time going back and forth every day, me explaining what I meant, him building it differently than I'd pictured it. Between all the hires, I spent somewhere around ₹5 lakh on engineering.
At the time, I thought that was the big mistake I made, but I don't think it actually was the first big mistake. The first big mistake was believing someone else could automatically build my product better than I could, just because of a resume.
Eventually there'd be a change I wanted, and I couldn't afford to keep paying for revisions. The engineer would basically say, "okay, no problem, contact me when you can, or "This is how it is supposed to be" So I'd go back to doing it myself. And when I did, I noticed something uncomfortable: some of what they built was technically fine, but it wasn't what we actually needed. I ended up redoing a lot of it myself anyway.
So the money didn't just disappear and It cost me time too you know like reviewing, explaining, correcting, rebuilding. And the worst part was realizing that some of what I'd paid for wasn't even something users had asked for. That one thing stuck with me. I'm glad I caught it very early instead of years and a lot more money later.
Money Makes It Easier To Move In The Wrong Direction
Before the grant, every feature I built cost me my own time, which meant I thought hard before starting anything. After the grant, that cost got less and less visible, because I could just pay someone else to do it. I think that changed how I made decisions without me really noticing at the time.
I had a lot of ideas, I mean a lot. I'd think "this would be useful," or "this would make the product better," or honestly, sometimes just "this would be cool to have" and then I'd build it, or pay to have it built.
Once I started paying closer attention to analytics and actually watching what people did instead of what I assumed they'd do, I realized a lot of what I thought was important wasn't important to our users at all. I'd built things that looked good on a dashboard, impressive when I looked at the product myself but users weren't touching them.
So I started removing things and I stopped assuming that because I personally thought something was useful, users would agree. Even now, one of the first things I do most mornings is make my tea and open analytics and look at what's working, what isn't, what users keep coming back to, what they ignore, what they would actually pay for. I started learning the product from users instead of just from my own head. What NexuSelf looks like today it is because of that shift, not out of the original plan.
The Food-Recognition Example
This is the clearest version of the mistake. Food recognition was very interesting to me. I could see why it might be useful. And because I had money, I had the ability to pay someone to build it. So I did.
The question I should have asked first was simple: do users actually need this? I didn't ask it seriously enough. It wasn't really a technical failure. The system itself worked, the mistake was that I built it without first knowing whether anyone actually needed it. I was checking whether I could build something, not whether I should.
The Wearable Request
There's one moment that changed how I think about the product itself,it is separate from the money. Most apps put things like connecting a wearable behind a Settings screen and users have to go find the toggle, flip it themselves. That's how I designed things too: if a user wants something, give them a screen to do it.
Then a user talked to Raha, our AI agent, and just asked, "can you connect my Whoop to this app?" They didn't want to go find the setting. They wanted to say what they wanted and have it happen. That sat with me for a while. Giving someone a setting doesn't mean they want to use a setting. Sometimes they don't want another screen or another button, they just want the thing done.
What's Different Now
If I receive the same grant again today, I'd spend it a lot more carefully than I did the first time I received it. If I were hiring right now, I wouldn't hire someone because their past projects look impressive. A polished, technically sophisticated project doesn't tell me whether someone can build what I specifically need. What I'd actually care about is how quickly they understand the problem and our project. They don't need to understand it better than I do, but they need to understand it well enough that we're actually solving and building the same thing, so I'm not explaining every decision from scratch.
I also don't think that money was wasted, exactly. It was an expensive lesson, but relatively cheap compared to what I've seen other founders spend learning the same thing. I caught it early. Thankfully I still had money left, and time to change direction and that's what I did.
The biggest shift is that I don't want to build things because they'll look good. I'd rather build something that looks unimpressive on a dashboard but actually helps someone, than something that looks great in a demo and goes untouched. I don't think I can write a line of code now without asking whether it adds friction for the user. If I think it does, I either don't build it, or I find another way.
I used to think capital gave me the ability to build more. Now I think it should give me the ability to learn faster. Those sound similar, but they're not.
I'm still a first-time founder. I'll probably make more mistakes. I just hope the next ones are cheaper.