Portrait of Dmytro Kuzmenko

Dmytro Kuzmenko

Assistant Professor @ NaUKMA·Research Fellow @ AIRlab, UniTo
CV Scholar GitHub LinkedIn X

I am a machine learning researcher working on embodied intelligence that is efficient enough to run under real compute budgets and trustworthy enough to act around people. My work spans world models and model-based RL, modular instruction-conditioned policies for multi-task robotics, and safety-aware robot behaviour in human-robot interaction.

I am an Assistant Professor at NaUKMA, where I teach machine learning, reinforcement learning, computer vision, and NLP, and run a student research program whose first student-led papers appeared at ECCV 2026 workshops. I am also a Research Fellow at AIRlab, University of Turin, working on assistive and socially intelligent HRI.

I received my PhD in Computer Science from NaUKMA in September 2026, advised by Dr. Nadiya Shvai, with a thesis on Resource-Aware Embodied Intelligence Through the Optimisation of Data, Compute, and Trust.

If you would like to discuss research, collaboration, or supervision, feel free to reach out by email.

News

Selected Publications

MoIRA preprint thumbnail
MoIRA: Modular Instruction Routing Architecture for Multi-Task Robotics
D. Kuzmenko, N. Shvai
Neurocomputing, 2026
Full paper arXiv 
@article{KUZMENKO2026132962,
title = {MoIRA: Modular instruction routing architecture for multi-task robotics},
journal = {Neurocomputing},
volume = {674},
pages = {132962},
year = {2026},
issn = {0925-2312},
doi = {https://doi.org/10.1016/j.neucom.2026.132962},
url = {https://www.sciencedirect.com/science/article/pii/S0925231226003590},
author = {Dmytro Kuzmenko and Nadiya Shvai},
}
EED Gym thumbnail
When Robots Say No: The Empathic Ethical Disobedience Benchmark
D. Kuzmenko, N. Shvai
ACM/IEEE International Conference on Human-Robot Interaction (HRI), 2026
Full paper arXiv Code Website 
@inproceedings{kuzmenko2026eedgym,
  title     = {When Robots Say No: The Empathic Ethical Disobedience Benchmark},
  author    = {Kuzmenko, Dmytro and Shvai, Nadiya},
  booktitle = {Proceedings of the ACM/IEEE International Conference on Human-Robot Interaction (HRI)},
  pages     = {168--176},
  year      = {2026},
  publisher = {ACM},
  doi       = {10.1145/3757279.3785547},
}
Proxemic risk assessment results thumbnail
Can Vision-Language Models Assess Proxemic Risk from Egocentric Robot Images?
V. Rudas†, D. Kuzmenko
EMR Workshop at the European Conference on Computer Vision (ECCV), 2026 (non-archival)
arXiv 
@misc{rudas2026proxemic,
                      title={Can Vision-Language Models Assess Proxemic Risk from Egocentric Robot Images?},
                      author={Vladyslava Rudas and Dmytro Kuzmenko},
                      year={2026},
                      eprint={2608.12515},
                      archivePrefix={arXiv},
                      primaryClass={cs.CV},
                      note={Accepted at the EMR 2026 workshop at ECCV 2026 (non-archival)},
                      url={https://arxiv.org/abs/2608.12515},
                }
Digital twin calibration pipeline thumbnail
Trajectory Design and Budgeted Querying for Digital Twin Calibration
V. Spitkovska†, D. Kuzmenko
CDEL Workshop at the European Conference on Computer Vision (ECCV), 2026 (non-archival)
arXiv 
@misc{spitkovska2026trajectory,
                      title={Trajectory Design and Budgeted Querying for Digital Twin Calibration},
                      author={Vladyslava Spitkovska and Dmytro Kuzmenko},
                      year={2026},
                      eprint={2608.08631},
                      archivePrefix={arXiv},
                      primaryClass={cs.LG},
                      note={Accepted at the CDEL 2026 workshop at ECCV 2026 (non-archival)},
                      url={https://arxiv.org/abs/2608.08631},
                }
TD-MPC-Opt preprint thumbnail
TD-MPC-Opt: Distilling Model-Based Multi-Task Reinforcement Learning Agents
D. Kuzmenko, N. Shvai
Preprint (under review), 2025
arXiv Code 
@misc{kuzmenko2025tdmpcoptdistillingmodelbasedmultitask,
                      title={TD-MPC-Opt: Distilling Model-Based Multi-Task Reinforcement Learning Agents},
                      author={Dmytro Kuzmenko and Nadiya Shvai},
                      year={2025},
                      eprint={2507.01823},
                      archivePrefix={arXiv},
                      primaryClass={cs.LG},
                      url={https://arxiv.org/abs/2507.01823},
                }

† Student I supervised

View all publications

Academic Service

Teaching

Assistant Professor, NaUKMA
13 theses supervised (12 BSc, 1 MSc)

Graduate courses:

Undergraduate courses: