• Designing new machine learning architectures to embed stylistic features into components of LLMs to determine predictiveness of a given piece of text compared to an author’s overall writing style

• Researching and developing novel stylistic embeddings that allow machine learning models to explicitly perform explainable inference when determining the source of an excerpt

• Assisted in creating lecture slides and learning objectives focusing on important topics that graduate students would need for their future careers in industry and research

• Drafting homework assignments and test questions that help graduate students learn the applied side of supervised and unsupervised learning

• Developed a machine learning model to accurately predict millions of missing internet traffic flow values for a given 24 hour period, allowing well informed traffic routing, significantly reducing the likelihood of congestion events globally

• Researched and designed forecasting models, custom validation metrics, and inference strategies with customers and stakeholders to create a financially efficient pipeline that helps downstream software optimize internet traffic routes

• Assisted in building a service that creates change log information used in routing calculations, reducing data staleness from 15 min to 1 min, decreasing overall latency experienced by all AWS customers

• Writing lesson plans and demos to convey introductory machine learning topics in a digestible format for over 200 undergraduate students

• Leading 4 machine learning projects for advanced students in text summarization, sentiment analysis, music, and music generation