Machine Learning

For this project, created for my final submission for a machine learning course I took at Harvard during graduate school, applied a series of generative adversarial neural networks to image datasets of Ed Ruscha’s iconic photographic series of Sunset Strip which span his streetscape photos from 1960-2007. Using the RunwayML platform, these image datasets used together with the ML models built into RunwayML generated a series of ‘feverish machine hallucinations’ depicting StyleGAN's reading/subsequent mimicking of Sunset Strip billboard text. Many thanks The Getty Image Archive, Stamen’s Shading Sunset image data portal, the RunwayML platform, and Ed Ruscha’s photographic work whose collaboration on this project allowed my final project for this class, DES3505: Machine and Image Processing to come to life.

I helped organize as part of my part-time work during graduate school, at Maxar, the SpaceNet.ai Challenge; a geospatial machine learning hackathon. SpaceNet delivers access to high-quality geospatial data for developers, researchers, and startups. Before SpaceNet, computer vision researchers had minimal options to obtain free, precision-labeled, and high-resolution satellite imagery. SpaceNet focuses on four open source key pillars: data, challenges, algorithms, and tools.

Developed a writeup developed on behalf of Maxar to elucidate how Maxar’s Deepcore.io platform machine learning capabilities could be used on used on behalf of the SpaceNet.ai Challenge. SpaceNet.ai baseline results run via DeepCore’s OpenSpaceNet platform and results were tested for validation and inference.

For my final project in Harvard course CS6483 I built a geographic visualization using a Natural Language Processing tool, Mordecai geoparser to detect every geographic place name ever mentioned in a Drake or Taylor Swift song. Using the Genius Lyric API, the geoparser ML model mordecai was applied to the entire cannon of Drake and Taylor Swift’s song archive.

Proud to have presented to UVA's Girls Who Code chapter as part of my role at Maxar. My presentation; An Introduction to Machine Learning was a part of UVA's Girls Who Code official yearly conference, Girls Hoo Hack.

As part of my professional work part-time at Maxar I helped organize and collate ML models and satellite imagery for the MBIE's Takiwaehere Geospatial Hackathon. The Hackathon took place over the weekend of 17-18 April, 2021 and was attended by 230 students across 68 teams at universities throughout New Zealand centered on geospatial machine learning exercises.