https://www.wiredamazon.comDeveloping a custom YOLOv11 AI model to detect Cocos from Castaña trees within drone images. Currently testing multiple model architectures and annotation strategies to identify the optimal model for high-precision detection. Project aims to establish a relationship between visible Cocos incanopy images and Castaña harvest volume, providing data-driven predictions for sustainable harvest forecasting.

• Conducted data collection using automated drone missions for Brazil nut tree analysis.• Trained a custom YOLOv8 AI model on 1,576 annotated images to achieve high evaluation metrics.• Implemented data augmentation techniques to enhance model robustness and accuracy.

Spatial and data analysis for forest carbon project development. Responsibilities include processing and analyzing large quantities of field data, analyzing land cover, utilizing forest growth and yield models, managing incoming inventory data, documenting work products, and assisting with project development.

Creation of Geomundo, an automated tool for batch editing satellite images and automated detection of illegal mining, illegal logging, road construction and deforestation using HR/VHR satellite Images.

• Conducted GIS and Spatial Ecology research for Spectacled Bear conservation using remote sensing with Landsat and Sentinel-2 imagery.• Utilized Maxent species distribution modeling to support conservation planning and habitat mapping.• Analyzed and predicted water hole locations in dry forests to aid in conservation efforts.

• Developed PyQGIS script to calculate GPS locations of Howler Monkey troops, automating calculations based on observer coordinates, bearing, and distance.• Digitized spatial layers to propose forest extensions for Howler Monkey habitats, identifying property owners for potential land acquisitions.• Analyzed roadkill data using PyQGIS to develop a predictive model for hotspots, recommending tunnels and overpasses.

The GetGo CarShare pilot project addressed the lack of effective alternatives for lower income persons living in affordable housing apartment properties to reduce car ownership costs. The Getgo carsharing model provided the most affordable and accessible form of carsharing to date. In close partnership with affordable housing property owners, GetGo placed cars in affordable properties in the Bay Area. Residents without credit cards or smartphones could have easy access to car rentals. Up to 500 lower income persons may save as much as $5,000/year in car ownership costs.

Provided Coordination in an affordable senior housing property between property management, development department, general contractor, moving company and building residents during a two year rehabilitation of the property. Responsibilities included weekly meetings to facilitate communication between the general contractor, property management, development department and building residents. Development and implementation of plans to deal with construction issues and building resident needs. Supporting residents and their families in accessing community resources and services. Advised and advocated for residents with emotional, physical and financial concerns related to the new building transition and construction. Utilized de-escalation techniques to diffuse conflicts between residents, families and property management staff.