Mexican PhD Student Wins Google Fellowship for Tiny AI

Luis Eduardo Garza Elizondo spent his childhood taking apart toys just to see what made them tick. Now that curiosity has led him to reshape the future of artificial intelligence.

The PhD student at Tecnológico de Monterrey is solving one of AI's biggest problems: its massive size. While most artificial intelligence relies on enormous data centers gulping electricity, Garza is teaching tiny microchips to think for themselves.

His invention, called TinyRL (Tiny Reinforcement Learning), lets devices learn and adapt without ever connecting to the cloud. These miniature systems can figure things out through trial and error, developing intelligence right where they sit.

In his university lab, Garza demonstrates with a small robot that starts completely clueless. It doesn't know how its wheels work or what its sensors detect, so it bumps into walls and spins aimlessly at first.

But after a few hours of tiny experiments, something remarkable happens. The chaos becomes coordination as the robot teaches itself to navigate, all without pre-programmed instructions or help from powerful computers elsewhere.

"You can literally see intelligence emerging from scratch," Garza explains. "We want to show that intelligence doesn't have to mean excess."

The Ripple Effect

This technology could transform how we live and work. Garza envisions factory robots that learn new tasks on the job, instead of needing complex reprogramming.

He imagines smartwatches that don't just track your heartbeat but actually predict health problems before they happen. Home assistive devices could adapt to their users' specific needs over time.

The environmental impact matters too. Today's large AI models leave enormous carbon footprints from the energy required to run massive server farms.

His work has caught Google's attention in a big way. The company selected Garza as one of just 255 doctoral students worldwide to receive its prestigious PhD Fellowship for 2025, which comes with funding, mentorship, and access to a global research network.

Garza leads Tec de Monterrey's Research Group for Industry 5.0, focused on designing technology that's smaller, smarter, and friendlier to both people and the planet. Multiple tiny processors can even work together, sharing discoveries like a networked ecosystem of intelligence.

That kid who couldn't stop opening his toys is now opening doors to a future where powerful AI fits in your pocket and learns from the world around it.