Starting and Growing a Career in Web Design

AI has made enormous progress in the digital world. But in physical businesses, operations remain complex, dynamic, and largely unmanaged by intelligent systems. Restaurants are just the starting point.

At Palona, we are building AI that operates in the real world: systems that can see, hear, and act across voice, vision, and workflows to improve real operational outcomes, in real time.

This requires a shift from predicting text to understanding how actions drive results. It means combining multimodal perception with causal reasoning, learning not just what is happening, but why it happens and how to improve it. Approaches like Joint Embedding Predictive Architecture point in this direction, but applying them to real-world operations remains an open challenge.

This is the context in which Hai Huang, PhD, joined Palona this week.

Hai spent over a decade at Google and recently co-authored research papers on JEPA with Yann LeCun, published on ICLR and ICML. His work has focused on building systems that unleash the full potential of JEPA beyond its traditional modalities, leveraging the semantic understanding necessary to ground complex decision making.

At Palona, he will focus on advancing our core intelligence layer — developing systems that learn from real-world interactions and continuously improve how businesses operate.

Why join Palona now?

“I wanted to work on AI systems that operate in the real world and have immediate impact. What stood out about Palona is the ability to take ideas from research and see them directly improve how businesses run day to day. There’s also a strong sense of mission — building technology that supports the people who keep these businesses running.”

We’re seeing this pattern consistently. People who have the option to work anywhere are choosing to work on problems that matter, with teams building toward something meaningful. A team came together based on a shared mission and purpose.

Looking ahead, we are continuing to invest in advancing how AI systems learn from and operate within real-world environments, extending beyond perception into understanding operations, financial dynamics, and the drivers of business performance.

We are still early, but the direction is clear. The next generation of AI will not just generate content. It will operate in the real world.

We’re building toward that future now.