A 15-year-old from Ontario built a bionic underwater “robot turtle” that detects what is quietly killing our seas and just won $50,000 for it

Most people see a turtle skim across a pond and move on. Evan Budz, a 15-year-old Grade 10 student from Burlington, Ontario, saw one and thought, what if a robot moved like that?

That question became BURT, short for Bionic Underwater Robotic Turtle, an AI-powered aquatic robot that mimics the swimming motion of a snapping turtle instead of the loud, turbulent propellers of most underwater drones. BURT has four flippers to move. The front two are for propulsion and the rear two are for steering and stability.

According to a ScienceDirect review on autonomous robotic systems for coral reef monitoring, the integration of visual data with AI and machine learning allows for real-time analysis of coral health, species distribution, and bleaching detection at scale. BURT is doing just that, but in a body that will leave the environment as undisturbed as possible.

Why it's important that it's quiet
Most underwater drones get the job done, but they add propellers, noise and physical disruption to the ecosystems scientists are trying to protect. Coral reefs, freshwater lakes and shallow coastal habitats are sensitive sites. A propeller-driven machine passing through them is like revving a motorcycle engine in a library.

BURT is about biomimicry, which is about building technology based on what nature has already worked out. Budz told Popular Science that when he saw the snapping turtle during a camping trip, he was captivated by how it moved through the water, fluid and non-disruptive. That was the design brief.

Published in Discover Applied Sciences by Springer Nature, the research highlights the growing demand for intelligent, maneuverable and low-impact underwater robotic systems and notes that bio-inspired designs provide a foundation for advanced marine surveillance and ecological monitoring. BURT fits right in with that direction.

What BURT can actually detect
A front-facing camera is on the robot, which is powered by a Raspberry Pi microcomputer. AI models are trained to detect signs of ecological trouble. For testing, Budz built simulated coral reef models and trained BURT to identify coral bleaching, invasive species, and plastic waste. In those tests under control, the detection accuracy was 96%.

Much of the testing was in his grandparents' backyard pool, just over 8 feet deep, before it moved to Lake Ontario. The real ocean is messier, murkier, and much less predictable, so those numbers will have to be validated on a larger scale. But 96% is no small thing for a working prototype built by a teen at Dr. Frank J. Hayden Secondary School in Burlington, using off-the-shelf parts.

BURT weighs about 11 pounds, is powered by a lithium battery for up to 8 hours, and has a solar panel to extend its operation. Its current cruising speed is around 0.5 miles per hour, which is about the speed a real sea turtle actually swims. That slower pace is not a disadvantage for conservation work; the most valuable thing is often steady observation over time.

The microplastics problem is next
The project has evolved quite a bit from its original form. Budz won the $50,000 Gordon E. Moore Award for Positive Outcomes for Future Generations at the Regeneron International Science and Engineering Fair in Phoenix, Arizona, in May 2026, where 1,727 finalists from 67 countries and territories competed. It is one of the most prestigious youth science awards in the world.

For that contest, Budz had transformed BURT into something new: a real-time microplastics detection platform with an AI-powered 3D holographic camera system. According to the Society for Science, which runs the Regeneron ISEF, his AI models were 94% accurate in identifying microplastics among other particles in underwater images, including pieces of plastic smaller than a red blood cell. He also received a Category Award in Environmental Engineering and two Special Awards. Prior to ISEF, BURT had already taken first prize at the European Union Contest for Young Scientists in Latvia in 2025.

Microplastics are some of the hardest environmental problems to track. They are small, they are everywhere, and traditional surveillance techniques are lagging behind. A peer-reviewed study published in the journal Sensors in 2024 developed an AI camera system to detect microplastics in real time and reported precision rates of 97% under laboratory conditions and 96% during field testing in a local river, demonstrating that computer vision-based detection is viable and scalable. Budz’s holographic camera system is moving in the same direction, bringing that kind of precision to mobile underwater monitoring.

Research on compact optical sensors and AI-enabled cameras on mobile platforms for real-time microplastic detection, published in Frontiers in Environmental Science, tested systems that continuously feed data to machine learning models and may enable near-real-time tracking of microplastics over large areas. BURT’s design is evolving to fit that exact use case.

Low cost, high ambition
One of the lesser-known but still important things about BURT is how it was built. Budz used easy-to-find, off-the-shelf parts. He didn’t need a university lab, a corporate budget, or special equipment. The robot has also been upgraded with front lights for murky water conditions and an ultrasonic transducer for obstacle detection, turning it into less of a science fair entry and more of a real research platform in development.

The oceans cover more than 70% of the planet’s surface and are under increasing threat from climate change, pollution and habitat loss. Current monitoring tools are expensive, intrusive or limited in scope. It’s not just a clever build; it’s a quiet, affordable, AI-equipped robot that moves like a turtle, can run for eight hours on a single charge and can spot a piece of plastic smaller than a red blood cell. It’s precisely the low-impact, high-precision instrument that marine scientists and conservationists across the US and elsewhere have been waiting for.