Jose Cols is a research scientist at the eScience Institute. His work focuses on natural language processing, large language models, and the software systems that support data-intensive research. He is especially interested in how multilingual models represent language across text and speech, and in making those models more efficient to use. With a background in production software engineering, Jose collaborates with researchers across disciplines to develop practical, scalable tools for working with language data. He also contributes to consultations, workshops, and learning programs supporting the UW research community.
Jose first joined eScience as a Text Mining Specialist, following several years as a software engineer building data platforms, information retrieval systems, and web applications. He holds an M.S. in Computational Linguistics from the University of Washington, an M.C.S. in Data Science from the University of Illinois Urbana-Champaign, and a B.S. in Computer Engineering from Andres Bello Catholic University.
Expertise:
- Multilingual NLP and machine translation
- LLM post-training and batch inference
- Information extraction and retrieval
- Distributed machine learning
- Scalable data pipelines
- Cloud infrastructure and DevOps
- Languages: Python, TypeScript, PHP, C++, R, and SQL