Computer Vision Development

During a internship at Scion, I helped developed computer vision annotations leveraging LINZ's regional aerial imagery for the detection of Radiata Pine, as well as focused annotations on multi spectral imagery for wilding pine detection, along side other GIS work. I also explored some of the YOLO models and more sophisticated implementations of Computer Vision and 3D Gaussians to explore tree stem texture extraction from non optimal imagery.

I've engaged in computer vision research and development through academic coursework and independent projects. Focused on exploring practical applications in image processing as well as developing the piplines to leverage computer vision models on data. I've explored developing models on my own drone imagery, but lack the quantity of data to really pursue any specifc projects. I have had success in developing models to explore what can be leveraged from the 0.075m LINZ imagery. Both of these projects have revloved around the detection of weeds in imagery.

I worked on developing and implementing Large Language Model inference systems for business analysis applications. I designed and developed the prototypes and inital production versions of automated workflows for performing a business analysis leveraring Flipviews existing platform to frame the anaylsis. This solution enabled faster onboarding and trailing of the site, as well as often providing left feild or otherwise overlooked aspects. It leveraged multiple different models to achive the best results as effiecently as possible.