I’m a second-year research-based MS student with the Computer Science department at CU Boulder. I did my undergraduate studies at Bowdoin College, where I received my BA in 2019. I majored in computer science, with a minor in music. In my time at Bowdoin, I studied machine learning, both as a research assistant on the KinoLab project, and for an independent study pertaining to machine learning backed baseball statistics. That work bred an interest in understanding how complex machine learning models could be applied to real-world scenarios, and made me appreciate that understanding a real world problem or dataset could be just as important as having technical machine learning skills to solve it.
Before coming to CU, I worked as a software engineer at ZoomInfo for about 4 years, where I developed and maintained large scale data pipelines. My work spanned a variety of fields, including productionizing proof-of-concept ML models, creating address normalization software, and designing/implementing significant components of a Scala Spark batch pipeline responsible for generating roughly 120 million entities.
At CU, I’ve become interested in geospatial machine learning research, particularly multi-modal foundation models. I’m interested in utilizing multi-task training, particularly masked auto-encoding and contrastive learning, to generate robust embeddings across a wide range of modalities. My hope is that such a model would be able to operate highly effectively with whatever data is on hand without requiring extensive pre-processing or inference to deal with different resolutions or missing data.
Outside of my research, I’m passionate about education. I’ve worked in outdoor education for ten years and recently taught Principles of Programming Languages for CU.
When I’m not busy with work, I’m an avid outdoorsman, part-time musician, cook, and baseball fan (lets go Mets!)