About

I'm Yekaterina Donegal — a data science graduate student at Pace University in NYC. My undergraduate background is in psychology and anthropology, which still shapes how I think about machine learning: I'm most interested in the human side of models, where the data comes from people and the predictions affect them.

Right now I'm a contributor on CAMELS / LSCA, a group capstone research project led with Watson Blair that aligns video, phoneme, and prosody modalities in a shared latent space for real-time human–AI interaction. Outside of research, I work on applied projects spanning retrieval-augmented assistants, time-series forecasting, and big-data pipelines.

A few things I care about

• Customer behavior, decision-making, and what data can (and can't) tell us about it.
• Multimodal models — combining language, vision, and audio rather than picking one.
• Tools that meet people where they already are (waiters, drivers, patients).
• Writing things down so the next person doesn't have to rediscover them.

Get in touch

Email contact@yekaterinadonegal.dev, or find me on GitHub and LinkedIn.