Welcome to my Project Collection#
Hi, I’m Spencer! This site is a curated collection of work that reflects my learning journey and professional experiences.
For featured technical projects across data science, each one includes context, methodology, and key outcomes to provide a clear picture of the problem-solving process behind it.
You can explore the full source code and documentation on my GitHub repository.
Featured Projects#
ETL for AI-Powered Messaging: Built data pipelines to extract structured insights from SEC 10-K filings, enriching company profiles for LLM-based sales messaging.
N-gram Pattern Mining in Menus: Mined over 30K dish names from restaurant menus using TF-IDF and machine learning to identify culturally meaningful compounds.
Enhancing LLM Visual Reasoning: Created an annotated dataset of visual reasoning tasks and fine-tuned models to simulate human-like problem-solving. Won Best Technical Innovation at UBC.
Multilingual Text Detoxification: Improved text detox systems for English and Russian using large language models, with measurable gains in style accuracy and content retention.
Other#
Brand Linguistic Insights & Research: A selection of anonymized deliverables from my time at Labbrand, where I conducted multilingual brand naming evaluations across markets.
Design Training & Portfolio: Selected projects from my Rhode Island School of Design training, showcasing skills in visual sensibilities, storytelling, and cross-disciplinary collaboration.