Katerina Margatina
Applied Scientist at Amazon Web Services (AWS)
United States
Summary
👤
Senior🎓
Top SchoolKaterina Margatina is an applied scientist and PhD researcher in natural language processing with eight years of experience bridging academic research and industry R&D at AWS and Meta. Her work focuses on active learning, data collection and evaluation, low-resource learning, uncertainty estimation, robustness, modeling disagreement, and fairness—areas that sit at the intersection of practical data pipelines and principled model evaluation. She has hands-on experience deploying research-driven solutions in production settings and teaching NLP courses during her doctoral studies at the University of Sheffield. A visiting PhD researcher in Copenhagen and an engineering graduate from NTU Athens, she brings a cross-cultural, systems-oriented perspective to NLP problems. Colleagues value her ability to translate nuanced research on calibration and disagreement into measurable improvements in data efficiency and model reliability.
8 years of coding experience
2 years of employment as a software developer
Visiting PhD student at the Department of Computer Science, Natural Language Processing, Visiting PhD student at the Department of Computer Science, Natural Language Processing at Københavns Universitet - University of CopenhagenCharles III University of Madrid (Universidad Carlos III de Madrid)Doctor of Philosophy - PhD, Machine Learning/Natural Language Processing, Doctor of Philosophy - PhD, Machine Learning/Natural Language Processing at The University of SheffieldElementary - Highschool, Elementary - Highschool at Arsakeia - Tositseia SchoolsDiploma (BEng&MEng), Electrical and Electronics Engineering, Diploma (BEng&MEng), Electrical and Electronics Engineering at National Technical University of Athens
English, Greek, French, Spanish
Github Skills (34)
transformers10
sgd10
natural-language-understanding10
deep-learning10
twitter9
python9
language-model9
lstm9
emotion-analysis9
active-learning9
monte-carlo9
transfer-learning9
transformer8
sentiment-analysis8
natural-language-processing8
Programming languages (2)
Github contributions (5)
katerinamargatina/katerinamargatina.github.io
Sep 2021 - Jan 2026
A beautiful, simple, clean, and responsive Jekyll theme for academics
Contributions:146 pushes in 4 years 5 months
responsive-jekyll-themeacademicscleanjekylljekyll-theme
mourga/transformer-uncertainty
Mar 2021 - Aug 2021
Code for evaluating uncertainty estimation methods for Transformer-based architectures in natural language understanding tasks.
Contributions:77 commits, 2 pushes, 1 branch in 4 months
natural-language-understandingtransformerdeep-learninguncertainty-estimationuncertainty-calibration
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.