Summary
Tatiana Shavrina is a research scientist manager with a decade of experience scaling NLP and multimodal R&D, currently leading research on Meta’s Llama team. She has deep industry leadership from roles at Snap and Sberbank where she built cross-functional AI orgs, shipped large-language-model products, and drove measurable user impact (e.g., dialog fine-tuning that boosted retention and a text-to-image skill that attracted millions of users). Tatiana combines rigorous academic training (PhD in Computational Linguistics) with hands-on engineering—having developed production NLP systems, corpora, and prototype platforms since early career work on corpus curation and spell-check/autocomplete. She excels at translating research into production roadmaps, managing multiple teams and release trains, and creating evaluation leaderboards that catalyze community model improvements. Known for pragmatic experimentation, she also maintains a research-minded developer’s curiosity, evidenced by early open corpus projects and sustained contributions to language-model evaluation and multilingual modeling.
10 years of coding experience
5 years of employment as a software developer
Master's degree Computational Linguistics, Master's degree Computational Linguistics at Higher School of Economics
Bachelor's degree Theoretical and applied linguistics, Bachelor's degree Theoretical and applied linguistics at Moscow State University
NYI Summer School Linguistics and Cognitive studies, NYI Summer School Linguistics and Cognitive studies at Stony Brook University
Turun yliopisto - University of Turku
UCREL Summer School in corpus-based NLP Natural language processing Big Data Data crawling, UCREL Summer School in corpus-based NLP Natural language processing Big Data Data crawling at Lancaster University
LxMLs summer school machine learning deep learning, LxMLs summer school machine learning deep learning at Instituto Superior Técnico
English, French, Armenian, Finnish