Hey all,

Happy Sunday from sunny Florida.

I came down here last week for my best friend’s wedding (my entire family went), and it was the most magical experience ever.

Now, let’s get into building revenue engines with marketing. What you’re about to read is super cool.

Amir Hanna is a mechanical engineer.

So, when I first learned he paused coding to engineer a UGC system that scaled his language learning consumer app, Parrot, to $1.5M ARR in 11 months with a top LTV:CAC of 6.4 (yes, real), Knight Vision invested and welcomed him into our portfolio of killer companies.

The best UGC systems are when brands ink long-term relationships with their top UGC creators, consistently test new hooks to find winners, and quickly produce new content.

Engineering these UGC programs is becoming increasingly more difficult.

More people are now doing them, which means there’s a higher volume of this content online, so everyone needs to be better to actually capture consumer attention.

Today, Amir is going to tell you exactly how he engineered a viral UGC program that took his app from bootstrapped → Y-Combinator → $1.5M ARR in 11 months.

I’ll let Amir take it from here….

15 months ago, I had no idea how online marketing worked. I thought selling products online was a hoax or a distant reality. 

I studied Integrated Engineering at the University of British Columbia. My only marketing experience was running my personal Instagram and a basic website. I also thought mobile apps were silly and not worth much attention.

That changed when my close friend Erik taught himself three languages in under a year by stitching together six different tools.

Around the same time, we realized something crazy: Duolingo doesn’t actually work for most people, despite having over 50 million daily active users.

Fast forward eleven months after launching Parrot.

We’re at:

We’re building TikTok-style language learning, starting with Spanish, and we were 100% bootstrapped before joining YC. Investors include: Head of Startups at OpenAI, Pioneer Fund, Phosphor Capital, Transpose Platform, Suhail Doshi (Co-founder of Mixpanel, $150M ARR), Alumni Ventures, David Parker, Julia Knight, and CEOs of several unicorns.

My co-founders, Julia and Erik, are software engineers and had no interest in marketing or growth, so that problem landed on me (lol).

By early 2025, one thing was obvious: writing code was getting faster and cheaper. Distribution was the bottleneck.

So, what does a mechanical engineer with zero distribution experience do?

I went all-in on TikTok.

I noticed people with no followers going viral, which completely broke what I thought I knew about influencer marketing.

TikTok felt like an underpriced asset. You could get real distribution without building an audience.

There are hundreds of billions of short-form videos watched every day.

The users I wanted were already there. 

So I created a new TikTok account for Parrot and committed to posting one video a day, improving one small thing each time.

The first 600 videos were bad. None went meaningfully viral.

What changed everything was how I approached understanding the algorithm like an engineer.

Social platforms make money by selling ads, so they optimize for watch time. If you don’t hook someone in the first three seconds, the video dies.

Once that clicked, everything else followed.

We built a repeatable system: find hooks that work, reuse formats, and turn views into product interest at scale, which has now amassed 110M+ VIEWS and over 200K downloads. 

That’s how growth engineering at Parrot started. Turns out engineering and growing TikTok pages are not too different. 

Everything above is context.

Here’s what actually worked + how you can copy it.

If you’re in a mature category like language learning, the data already exists.

I would:

This shows you exactly what’s working: hooks, text, formats, and angles. You can adapt any of these winning videos to include your product. 

The hard part is subtly showing your product without making it feel like an ad. 

One of our best-performing videos came from this exact process.

The hook: “when I realized the best way to learn Spanish is the same way a child learns” kept working over and over. 

Filmed casually at a laptop. The app appears briefly, almost incidentally.

It works because it feels like a recommendation from a friend, not a sales pitch. User-generated content feels native. That’s why it converts.

Once you know what works, volume matters.

To do that, you need creators. We source them in three main ways:

If you don’t want to manage human creators, you can experiment with AI UGC or slideshow content. My take: people are starting to recognize AI-generated talking videos, but AI slideshows still work well. This will likely improve over time.

There are a few companies building phone farms. These phone farms will scroll like a human, like videos, and post videos as well with proxies.

To programmatically mimic human behaviour and avoid bot detection from TikTok and Instagram.

The idea here is that you can efficiently post 1,000 pieces of content every day and maximize the reach of the content + your learning. 

We built a structured onboarding program called Parrot Flight School. The program includes:

Once you have enough volume, you can:

Last thing:

None of this matters if your product isn’t good or if people don’t care. I probably should’ve led with that.

The path to building a billion-dollar consumer company should also involve viral loops and network effects. A way to acquire customers organically with $0 spent on marketing. 

Product-led growth is the best form of growth. It’s also the most difficult to crack. 

At Parrot, we still have a long way to go. But we’re on track to build the last language-learning tool people ever need, and to bring it to millions around the world. A while taking down Duolingo. 

If you want to speak Spanish fast, just by scrolling videos, Parrot has helped thousands of people get there.

Did you guys like this guest spot from Amir?! Crazy successful, tactical story.

UGC systems that drive quality traffic are brutally hard to build. Amir figured it out with a mechancial engineering background and used it to scale Parrot. Impressive, killer energy.

I had to get all the tea for you guys. So, here it is.

I hope you have an incredibly productive week ahead.

Julia