profile_picture
Evgeniia Tokarchuk

Hi there. I am Evgeniia Tokarchuk. I am NLP and machine learning researcher specializing in representation geometry, neural machine translation, retrieval-augmented models, and efficient modeling. My PhD was completed at the Language Technology Lab, University of Amsterdam under the supervision of Dr. Vlad Niculae, whom I was honored to work with.

Before that, I recieved my master’s degree at RWTH Aachen University, where my research path began at the Human Language Technology and Pattern Recognition Group. My master thesis “Tight Model Integration for Pivot-based Neural Machine Translation” was the result of a collaboration with eBay Aachen. Along the way, I was fortunate to work with great mentors and collaborators.

Outside of research, I enjoy reading and traveling. I am also a member of the Dutch Haedong Kumdo Association and I practice Haedong Kumdo Jinyoung Ssangkum Ryu at Dojo Den Drijver in Amsterdam. In addition, I am involved in local volunteering initiative aiming to support Ukraine πŸ‡ΊπŸ‡¦

Feel free to contact & connect!

News

Recent Publications

Representation Collapse in Machine Translation Through the Lens of Angular Dispersion, 2026, Findings of the Association for Computational Linguistics: EACL 2026
Evgeniia Tokarchuk , Maya K. Nachesa , Sergey Troshin , Vlad Niculae
Angular Dispersion Accelerates k-Nearest Neighbors Machine Translation, 2025, Findings of the Association for Computational Linguistics: EMNLP 2025
Evgeniia Tokarchuk , Sergey Troshin , Vlad Niculae
Keep your distance: learning dispersed embeddings on π•Šm, 2025, Transactions on Machine Learning Research (TMLR)
Evgeniia Tokarchuk , Hua Chang Bakker , Vlad Niculae
The Unreasonable Effectiveness of Random Target Embeddings for Continuous-Output Neural Machine Translation, 2024, NAACL
Evgeniia Tokarchuk , Vlad Niculae
On Target Representation in Continuous-output Neural Machine Translation, 2022, RepL4NLP 2022
Evgeniia Tokarchuk , Vlad Niculae
Integrated Training for Sequence-to-Sequence Models Using Non-Autoregressive Transformer, 2021, IWSLT 2021
Evgeniia Tokarchuk , Jan Rosendahl , Weiyue Wang , Pavel Petrushkov , Tomer Lancewicki , Shahram Khadivi , Hermann Ney
Investigation on Data Adaptation Techniques for Neural Named Entity Recognition, 2021, ACL SRW 2021
Evgeniia Tokarchuk , David Thulke , Weiyue Wang , Christian Dugast , Hermann Ney
Towards Reinforcement Learning for Pivot-based Neural Machine Translation with Non-autoregressive Transformer, 2021, Reinforcement Learning for Real Life Workshop
Evgeniia Tokarchuk , Jan Rosendahl , Weiyue Wang , Pavel Petrushkov , Tomer Lancewicki , Shahram Khadivi , Hermann Ney