DeepWeeds edge classifier
On-device weed classifier for the Australian DeepWeeds dataset — a 4.2M-parameter MobileNetV3 that beats the paper's ResNet-50 baseline, with ONNX/INT8 export and edge benchmarks.
Software engineer · Brisbane, Australia
I work on edge AI and computer vision for UAVs — PyTorch and ONNX models optimised to run on small hardware, and PX4/ROS 2 flight systems to carry them. Before that, a decade of commercial mobile development, including iOS team leadership. Currently completing a Master's in AI & Robotics at QUT.
On-device weed classifier for the Australian DeepWeeds dataset — a 4.2M-parameter MobileNetV3 that beats the paper's ResNet-50 baseline, with ONNX/INT8 export and edge benchmarks.
iOS app that turns shared links into clean recipe cards — on-device AI extraction, meal planning, and private iCloud sync.