Maintenance engineers in the transportation industry wanted to use a neural network to classify ultrasonic scans. They were concerned about connectivity and latency and wanted to run their model in an edge device.
I deployed an optimized neural network on the NVIDIA Jetson, ready to take ultrasonic scan data and classify defects.
Complete with a custom 3D printed case.
With a 5 hour battery life and 2500 fps inference speed, my client is now able to complete inspections faster and more thoroughly.
“Steel's ability to tackle complex tasks independently was remarkable and I have come to trust him for challenging problems.
He has also been a great communicator providing consistent and proactive updates. Highly recommend him for any complex technical needs!”
-Amit Bohara, Founder, Altosphere