Geospatial Analysis and Algorithm Development

I amcarolyn-headshot an applied mathematician and developer, with 20 years of R&D experience in the mapping and remote sensing industry. I develop algorithms and systems for extracting information from imagery, producing map data, and improving the accuracy of maps produced with the aid of remote sensing imagery.

My work history includes:

  • 5+ years with DigitalGlobe, the industry’s leading producer of high-resolution, high-geolocation-accuracy satellite imagery, working on geolocation improvement and image information extraction problems;
  • 4 years with Microsoft working in location-based services, developing streetview geolocation algorithms and solving text-based location problems;
  • 10 years with Vexcel Corporation working on SAR geolocation and image information extraction algorithms.

I’ve developed a scalable solution for detecting human settlements in hundreds of thousands of square kilometers of satellite imagery. I’ve developed new approaches to cleaning, normalizing, and deduplicating local search data. I’ve developed an automated, scalable solution for improving the geospatial accuracy of commercial satellite imagery, using large-scale photogrammetry. I helped invent the algorithms that were used to edit the SRTM digital elevation models in record time. Please see my Projects page for more details.

I speak the languages of business, algorithms, computers, and mapping. I pride myself on my ability to explain technical matters clearly and at every level, in both written and spoken form. I have managed teams of individuals tasked with developing production systems on time and under budget.

I understand satellite, aerial, and ground imagery. I understand lidar, GPS, synthetic aperture radar (SAR), and GIS. I code in C/C++, Python, and R, and my work utilizes the open-source geospatial tool stack including PostGIS, QGIS, and GDAL/OGR/Proj4.

I have studied and applied mathematical and machine learning methods including (but not limited to) nonlinear regression and optimization, generalized linear models, SVMs, random forests, probabilistic graphical models, and Bayesian models.

I’ve watched the location technology industry grow from data-starved to overwhelmed with data, and I’ve kept up with it all the way. I’m fascinated by the interesting problems that arise in this industry, and by the inventive ways in which they’re being solved.

Please contact me with questions about how I can help you to solve your algorithmic and mapping problems.


(Image: Landsat Image of the Buba River in Romania, courtesy US Geological Survey).