Maksym Tarnavskyi
Team Lead NLP Data Scientist
Poland
Summary
👤
Senior🎓
Top SchoolMaksym Tarnavskyi is a Team Lead NLP Data Scientist based in Poland with eight years of commercial experience building and shipping production NLP systems. He progressed from intern to team lead at Shelf, leading cross-functional projects and driving model-to-production improvements since 2022. His work spans tokenization, indexing, and pipeline efficiency—highlighted by contributions to the well-known GECToR grammatical error correction project where he implemented batch tokenization and added LLM support. Maksym pairs rigorous academic training in data science and systems analysis with hands-on ML engineering, focusing on making NLP components faster and more compatible with modern tooling. Colleagues describe him as a pragmatic leader who blends research-aware approaches with practical refactors that reduce technical debt.
8 years of coding experience
5 years of employment as a software developer
Master of Science - MS Data Science, Master of Science - MS Data Science at Ukrainian Catholic UniversityIgor Sikorsky Kyiv Polytechnic Institute
English, Ukrainian, Russian
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (9)
transformers10
tokenize10
tokenizer10
nlp10
python10
bert10
natural-language-processing9
sequence-labeling9
roberta8
Programming languages (4)
Github contributions (5)
grammarly/gector
Jul 2021 - Aug 2021
Official implementation of the papers "GECToR – Grammatical Error Correction: Tag, Not Rewrite" (BEA-20) and "Text Simplification by Tagging" (BEA-21)
Role in this project:
ML Engineer
ML EngineerContributions:18 commits, 3 PRs in 10 days
Contributions summary:Maksym primarily focused on enhancing the tokenization and indexing functionalities within the grammatical error correction project. Their contributions include implementing batch tokenization using fast tokenizers, refactoring the tokenization process, and adding support for large language models. These changes involved modifying core files related to tokenization, suggesting a focus on improving the efficiency and compatibility of the project's NLP pipeline. The user also addressed dependencies by incorporating legacy components.
grammatical-error-correctionnlptransformersxlnetbert
MaksTarnavskyi/ml-production-course
Sep 2022 - Nov 2022
Homework tasks for Prjctr Machine Learning in Production course
Contributions:5 reviews, 26 PRs, 116 pushes in 2 months
productionmachine-learningin-production
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.