Every so often, someone asks me the same question.

You were the CTO at Tinder. Why didn't you just build another dating app?

I understand why they ask. I joined Tinder in 2016 after years at Yahoo, where I was a VP of Engineering. At the time, the move puzzled people. Yahoo was large and established. Tinder was smaller, scrappier, and carried plenty of negative press. From the outside, it looked like a step down.

To me, it was never about company size. It was about the problem.

What Actually Stayed With Me

Tinder looked simple. A swipe, a match. But underneath that simplicity was a massive technical and human challenge. We were building something that had to work across countries, languages, behaviors, and emotions, and it had to feel effortless while doing it. That is usually the hardest kind of technology to build.

Last week, I was at a business dinner when someone mentioned he met his husband on Tinder. I have heard versions of that story many times. I never get tired of hearing them, because they remind me why building technology matters.

Technology often gets measured by revenue, valuation, or downloads. Those things matter, and Tinder became one of the highest-grossing apps in the App Store during my time there.

But what stays with me most are the people. Millions of introductions. Families that exist today because two people swiped right and crossed paths.

That is the privilege of building products.

You may never meet most of the people whose lives are changed by what you build. But every so often, someone sits across from you at dinner and tells you their story, and all the late nights and hard decisions become very real.

Why I Did Not Build Another One

Building another dating app would have been the easier path. I know the category, the growth mechanics, the trust and safety challenges. But I have never been motivated by repeating what I already know. I am motivated by learning what I do not.

That mindset has guided most of my career, from Yahoo to Tinder to Meta to Google and now Palona. The most rewarding work has never been following an existing roadmap. It has been helping create one.

The next challenge, as I see it, is generative AI that can actually complete complex tasks across modalities. AI that can listen, see, reason, remember, and adapt, not just respond to a prompt in a text box. AI that understands what is happening in the real world, not only what a user types into it.

That is what we are building at Palona. And it starts with a question bigger than any one industry. What happens when AI moves beyond screens and starts helping physical businesses operate better?

Why We Started With Restaurants

Most software over the past decade has lived in digital environments. Feeds, ads, dashboards, online transactions. Those systems are powerful, but they are also relatively contained. The digital world gives you clean data and predictable environments.

A restaurant is nothing like that.

Guests walk in. Phones ring. Orders change. A table waits too long. A kitchen gets overwhelmed. A manager makes hundreds of small decisions a day, often with incomplete information, and there is no pause button to debug any of it.

That is exactly why restaurants interested me. They are among the most operationally demanding businesses on earth, and every minute matters.

Most restaurants already run on a stack of systems, a POS, a reservation platform, a phone line, a handful of delivery apps. But those systems usually capture fragments of reality. None of them actually understand what is happening across the whole business.

What We Are Actually Building

Palona is built to close that gap. We are building multimodal AI agents that understand the physical world well enough to help real businesses operate with more intelligence and more care.

Voice, vision, and operational intelligence are not separate products in our view. Together, they form the foundation of something closer to an AI operating system for physical businesses, one that can observe what is happening, understand the context behind it, and help teams make better decisions in real time.

This work is hard. It is hard technically, and it is hard because the real world is messy and restaurants are not controlled environments. There is no established playbook for this yet. Honestly, that is exactly why I wanted to work on it. I am drawn to problems that are early and not yet fully understood, the kind where conviction matters but curiosity matters even more.

At Tinder, the question was whether technology could help people connect. At Palona, the question is whether AI can help real-world businesses operate with more intelligence. Different problem. Same instinct that has guided me through every chapter of my career.

I still smile every time someone tells me they found the love of their life on Tinder. But I am equally energized by a different kind of story now. A restaurant owner who finally has visibility into their operation they never had before. A team that can serve more guests without burning out. A guest who feels remembered and cared for because the business finally has the intelligence to deliver that experience consistently.

Different stories. Same reason to build.

If you're curious what real-world AI could look like in your restaurant, book a call with our team. We'd love to show you