UCLA Computer Science Student, Winner of Google AI Contest, Dreams Big of Building More Intelligent Tools

Jonathan Ouyang recalls his first memory of programming, learning how to draw pixel art by writing basic code in the first grade. Now, more than a decade later, the rising second-year UCLA computer science student has built a program that recently won him the honor of Best Overall App in the Google Gemini API Developer Competition — a global contest for artificial intelligence developers using Google’s latest multimodal model.

Ouyang’s project, “Jayu”, is an AI assistant that lives directly within the user’s desktop environment. Rather than opening a chatbot window or copying and pasting prompts, users can summon Jayu with a voice command and interact with it in real time without needing a separate browser window.

The project stemmed from Ouyang’s frustration with having to switch tabs every time he needed help from AI. By integrating the assistant directly into the user interface, Jayu is designed to enable a smoother and more seamless experience. The AI assistant is also able to understand a user’s hand gestures, such as scrolling up or down on a screen without touching the keyboard. 

“Dream bigger,” said Jonathan Ouyang. “You can build so many cool things with AI. I think anything can be possible if you can just dream it.”

Building Jayu was Ouyang’s first time creating a complete software project from start to finish. Until then, most of his programming experience came from coursework or modifying existing code. He entered the Gemini competition in 2024 as a freshman at UCLA after seeing ads from Google on social media.

“I thought this would be a good opportunity to test my own abilities and prove to myself that I was worthy of being accepted to a school like UCLA,” he said. “I never dreamt in a million years that I would be lucky enough to win a competition of this scale.” 

But the process wasn’t easy. The greatest challenge, he said, was getting the Gemini model to follow instructions reliably. That meant spending countless hours refining prompts through trial and error, trying to discover which phrasing the model would best understand. 

“There’s no secret technique,” he said. “It’s like trying to give instructions to a 10-year-old child. Some things just work and some things just don’t, and there isn’t always a clear reason as to why.”

The name Jayu — Korean for freedom — is both a nod to Ouyang’s Korean heritage and a reflection of the platform’s purpose and technical impact. Ouyang said it gives users freedom from the hassle of toggling windows while giving the Gemini model an unparalleled level of flexibility and access.

Google for Developers

Jayu, the AI voice assistant, responds to user commands by recognizing hand gestures.

Growing up in the heart of Silicon Valley, Ouyang was immersed in technology from an early age. His hometown of San Jose, a hub for major tech companies, made working with technology feel like a natural choice.

“Being surrounded by all these people in tech and living in an area with all of the big tech companies, you hear about all these new technologies every single day,” he said. “Having tech be such a prevalent part of my life influenced me to go into this field.”

That interest was reinforced at home. With a family lineage of engineers going back to his great-grandfather and parents who both earned master’s degrees in electrical engineering, Ouyang was enrolled in programming and robotics camps as early as elementary school. He recalls being introduced to Python, Java, app development and even artificial intelligence long before college — though his real fascination with AI didn’t surface until he joined an AI lab at San Jose State University as a junior in high school. Ouyang said he found the experience incredibly interesting and exciting despite having no background in the field at the time.

At UCLA, Ouyang has continued to explore the frontiers of AI, and developed an interest in robotics as well. He conducts undergraduate research in the UCLA Robot Intelligence Lab, led by Yuchen Cui, an assistant professor of computer science at the UCLA Samueli School of Engineering. He is also involved in research with Stanford’s Intelligent and Interactive Autonomous Systems Group in collaboration with the Toyota Research Institute, focused on shared autonomy in driving.

Ouyang said he was inspired to explore robotics by the advances he saw in robots in the world around him and that his experience in Cui’s lab has been especially formative.

“We used to program robots with very specific behaviors, like following a line or picking up traffic cones,” he said. “But now, with the rise of AI, we can have robots doing things that were never thought to be possible just 5 years ago.”

Outside the lab, Ouyang also serves as operations director for the Association for Computing Machinery’s UCLA chapter and works as a software engineer at the Daily Bruin. These experiences have helped him apply his technical skills to a broader campus community.

Ouyang’s long-term goal is to continue working at the intersection of AI and robotics, with plans to pursue a Ph.D. in the field. He says he hopes to keep building intelligent agents like Jayu while pushing the boundaries of how humans and machines interact.

When asked what advice he would give to other students aspiring to compete in AI challenges, his message is ambitious yet accessible.

“Dream bigger,” he said. “Move away from the idea that AI is just text-in-text-out chatbots. You can build so many cool things with AI, and with the tech just getting better and better and the speed that AI is advancing, I think anything can be possible if you can just dream it.”