Octavia Sulea
AI researcher · San Francisco
I work on natural language processing, multimodal systems, and model evaluation. My research spans language learning, legal prediction, speech, and semantic compression.
I have a Ph.D. in Computer Science from the University of Bucharest. My work has taken me through DFKI, PARC / SRI, and Hypernym, and into applied ML projects with Lark Health, GumGum, and Struck.
Selected papers
- Hypernym Mercury: Token Optimization through Semantic Field Constriction and Reconstruction from Hypernyms Chris Forrester, Octavia Sulea · arXiv, 2025
- MorphoGen: Full Inflection Generation Using Recurrent Neural Networks Octavia-Maria Sulea, Steve Young, Liviu P. Dinu · CICLing 2019
- Unsupervised Inflection Generation Using Neural Language Modeling Octavia-Maria Sulea, Steve Young · IWANN 2019
- Exploring the Use of Text Classification in the Legal Domain Octavia-Maria Sulea et al. · ASAIL 2017
- Predicting the Law Area and Decisions of French Supreme Court Cases Octavia-Maria Șulea, Marcos Zampieri, Mihaela Vela, Josef van Genabith · RANLP 2017
- Using Word Embeddings to Translate Named Entities Octavia-Maria Șulea, Sergiu Nisioi, Liviu P. Dinu · LREC 2016
More papers: Google Scholar · dblp · ACL Anthology
Invention
System and method for change point detection in multi-media multi-person interactions Co-inventor · Xerox Corporation US 12,744,037 · Granted September 22, 2026
Interviews & talks
- Octavia Sulea Is the Femme Future of AI and Data ScienceAudiofemme · 2020
- Data Augmentation for Sequence Labeling: A Case Study in Food ParsingNLP Summit · 2021
- Identity in the age of AIInterview with Alan Cruz · 2025
Projects in the news
Coverage of projects I’ve worked on.
- Struck’s dating app for the Co–Star crowd hits the App StoreTechCrunch · July 2020
- Struck, an astrology-based dating appLos Angeles Times · September 2020
- Lark Health’s approach to managing chronic diseaseBusiness Insider · October 2020