Their words have been edited for length and clarity.Praneet Dutta: I met Joe on my first day at Google years ago. I was leaving orientation when we ended up in the same Uber Share ride.Joe had joined Google only a few months prior. We were not on the same team; I was in machine learning, and he worked in cloud infrastructure, but over the next few months, we realized we had similar career aspirations.Joe Cheuk: That was roughly eight or nine years ago that we met.

I stayed at Google for around two years and then moved on to Meta, where I worked on advertising infrastructure. Most recently, I worked at Databricks on trust and safety.Dutta: After two years working in machine learning at Google Cloud, I moved over to Google DeepMind, where I stayed until August 2025. Joe and I always kept in touch as friends, but also with the implicit understanding that we wanted to start a company together someday.We finally left our respective roles at the same time last year to cofound our AI startup, Pomo.

Business Insider wants to talk to founders of AI-powered companies and teams with fewer than 10 workers, and employees working alongside AI agents and bots, to understand what the "Tiny Teams" era really feels like — the wins, the fears, and the human skills that stand out. Share your story by filling out this quick form. We always wanted to start a company together, but waited years for the right momentCheuk: Even though we were on different teams at Google, we found ways to collaborate on different side projects with each other.

I was very interested in learning more about machine learning, and he was interested in the infrastructure side.We had many conversations about whether the Big Tech environment was actually the best fit for us, because we wanted more freedom to make decisions and move quickly. We both realized that a startup environment could be a better fit.Dutta: When Joe moved to Seattle, he still visited, and we kept in touch. We both were exposed to the advertising and marketing ecosystem through our previous roles.

His exposure was with Meta, and I co-authored research with my team at Google DeepMind to improve image generation for advertisers' products.Last year was a pivotal point because that's when the timing aligned, and we saw our opportunity to use AI to help companies make marketing decisions faster.The timing was finally right, and AI made our ideas possibleDutta: Early 2024 was a paradigm shift with what we saw these AI models were capable of. We also started interviewing marketers, and those interviews made us realize: the technology is now there, the opportunities are there, the Silicon Valley funding lifecycle has always been there, but now it's much faster.We also both secured our green cards, which gave us the immigration stability we needed.

All of these things combined made us feel that, OK, now's the time to be doing this. My last day at Google was bittersweet, and I'm still friends with some of my old colleagues. We often miss Big Tech, but we know that to get great things done, you have to be uncomfortable sometimes, and we understand the risks we're taking.The biggest shift from Big Tech was how fast we have to make decisionsDutta: The fundamental difference going from Big Tech to a startup comes down to decision-making speed.

The cost of indecision for us is much, much higher as a startup. Joe and I, as co-founders of the company, need to always be thinking about how quickly we can move the business forward.In Big Tech, there are many different layers of approval. This sort of culture shift to immediate proactivity has been the biggest for us to get used to.Cheuk: The reverse of this, though, is that all these processes and solutions that we have seen in Big Tech help us not second-guess ourselves as much when we need to move quickly.

Now, if I'm looking at a similar problem or process, the solution just pops into my mind.There's no reason for us to slow down and think too much. It's been validated as the right choice by so many engineers who are smarter than me because I've watched them do it. As long as it's reasonable and not hard to build, we just go ahead and do it the way it's already being done.Grind mode is the most valuable skill right nowCheuk: Having the willingness to grind is a very important factor right now. We think that's one of the most crucial things in the AI