From open courseware to systems in the field
Roberto began studying AI in 2012, working through MIT and Stanford open courseware in the same year AlexNet helped change the field. By 2016, he was publishing brain.js experiments with character-level neural networks generating their own answers.
In 2018, he built a TensorFlow flood-prediction network using Houston stream-gauge data. His work has included updating the model on a new storm in about two hours. In 2020, he earned a Microsoft certification in reinforcement learning.
Today, his ESP32-CAM field units run a 4 MB, int8-quantized YOLO model, detecting his fine-tuned classes and sending detections home over LTE. Through AI-Rey, he turns AI research into practical business applications.
A willingness to act
As a young lifeguard, Roberto faced a rescue far beyond the pool. A girl fell from a cliff into a lake, broke her arm, and had no way to get out. He jumped in after her and swam hundreds of feet through open water, carrying her to shore.
Engineering started in the machine shop
While still in high school, Roberto machined precision parts for medical devices on CNC mills. He also placed on the speed-drafting podium two years running while keeping up with his schoolwork.
He earned a Bachelor of Science in Mechanical Engineering Technology, studying machine design, thermodynamics, and materials. A summer internship at a New Mexico potash mine brought him into the surface operations, where large mechanical systems and industrial chemistry turned raw material into finished product.
Beyond the workbench
Outside work, Roberto is a Masters-level StarCraft II duos player and a rock climber who climbs 5.12c. Strategy, persistence, and solving difficult problems are part of how he spends his free time, too.