In August 2025, I joined Kable.tv on a mission to build ambient AI—something closer to the real Jarvis from Iron Man than another assistant waiting inside an app.
For AI to feel truly ambient, it needs to understand more than the words typed into a box. It needs to hear, see, and respond to what is happening around it. That makes audio and video foundational to the experience, not features added later. Real-time media has been my area of expertise for more than a decade, so the opportunity felt like a natural extension of the work I had already spent much of my career doing.
It was also a chance to build a product with AI at its core. I did not want to add an AI feature to an otherwise conventional product. I wanted to work on the harder question of what becomes possible when intelligence shapes the product from the beginning—its interface, its architecture, and the way people relate to it.
Building for the home makes that question much more interesting. A household is not a single user with a single account and a private screen. It is a shared space with different people, relationships, routines, and expectations. An ambient system has to understand when it should help, when it should ask, and when it should stay out of the way.
That also makes privacy a product problem, not just a policy or security problem. A useful ambient assistant needs enough context to be helpful, but the most personal context in someone’s life lives inside their home. Earning the right to be present there requires deliberate choices about what the system perceives, what it remembers, who can access it, and how clearly people can understand and control those boundaries.
The people made the opportunity even more compelling. Kable.tv brought me back together with people I had worked with at Hopin: founder Johnny, Math Cassagnes, and Tim Chang. Starting something ambitious is difficult enough; getting to do it with people I already knew and trusted made the leap easier.
There was one more part of the opportunity that made it feel unusually personal. Johnny wanted the team together in San Mateo, California. My wife has family there, and he was open to sponsoring our move to the United States. The location was not incidental to the role—it could bring our family closer to people we care about while giving the team the shared environment the mission called for.
That possibility arrived almost a decade after I first had the chance to move to the United States for work and turned it down to accept a Fulbright scholarship. At the time, I chose education and trusted that the professional door might open again. I wrote more about that decision in Choosing Fulbright when another door opened. Joining Kable.tv felt like that door returning in a form I could not have planned: the right technical problems, people I was excited to work with again, and a move that made sense for my family.
The mission is ambitious. Making AI feel present without making it intrusive will require solving difficult problems across real-time media, intelligence, product design, and privacy. Those are exactly the kinds of problems I wanted to work on—and why I joined.