kyry.me/cv

Curriculum vitae

download pdfkkolesni@kent.eduGitHubLinkedInX

Education

B.Sc. Information Systems and Cyber SecuritySep 2023 – Jan 2027 (expected)Vilnius University, Institute of Social Sciences and Applied Informatics, LithuaniaGPA 3.8/4.0. Merit scholarship, five times.
Exchange semester (ISEP)Spring 2026Kent State University, Kent, OH, USAGPA 3.8/4.0. Dean's List.
Exchange semester (ISEP)Fall 2025University of Vermont, Burlington, VT, USAGPA 3.8/4.0.
Exchange semester (Erasmus+)Spring 2025Tampere University, Tampere, FinlandGPA 4.67/5.

Research experience

Robotics ResearcherMar 2026 – presentAdvanced Telerobotics Research Lab, Kent State University (advisor Jong-Hoon Kim), Kent, OH, USA
  • Built and open-sourced a video-to-humanoid pipeline that turns one monocular runway clip into a walking policy: 3D motion recovery (GVHMR), retargeting (GMR), motion-tracking RL in Isaac Lab (BeyondMimic), MuJoCo sim-to-sim validation, then the real Booster K1.
  • The learned catwalk gait finished 20/20 consecutive 5 m hardware trials with no falls and walks with crossover steps: step width −0.8 to 1.8 cm against 5.1 to 11.4 cm for the stock controller (IROS 2026 Late Breaking Results, first author).
  • Stylized and teleoperable locomotion for the K1 in mjlab: a catwalk style with adversarial motion priors, a legs-only policy that leaves the arms to a teleoperator, and walking policies fine-tuned for 3D-printed heel shoes.

Industry

Software Engineering InternMar 2025 – May 2025SoftServe, Remote
  • Developed user authentication and a recommendation module end to end, from React components to Node.js REST endpoints and PostgreSQL queries.
  • Designed relational schemas and integrated REST APIs across the stack; every change went through code review and tests before merge.

Publications

  1. Kyrylo Kolesnichenko, Irvin Steve Cardenas, Jong-Hoon Kim. “Learning Expressive Humanoid Locomotion from Monocular Runway Videos for Robot Fashion Shows.” IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2026), Late Breaking Results (LBR), 2026, presented, poster. arXiv:2609.27003
  2. Kyrylo Kolesnichenko, Martas Ambraziunas. “Can Structured Prompting Overcome Knowledge Bias in Vision-Language Models?” Proceedings of the 31st International Conference on Information Society and University Studies (IVUS 2026), CEUR Workshop Proceedings, 2026, presented, to appear.
  3. Raiyan Ashraf, Yanni Liu, Kyrylo Kolesnichenko, Sruthi Ganji, Jong-Hoon Kim, Ju Yup Lee. “Can Motivational Interviewing Guided Dialogue Improve Empathy in Human–Robot Interaction? A Controlled Study with Baseline Comparison.” International Conference on Intelligent Human–Computer Interaction (IHCI 2026), 2026, accepted.

One paper at IEEE ICRA 2027, under review.

Projects

Expressive Motion2026 – 2026-09From a monocular runway video to a catwalk gait on the Booster K1.
Catwalk gait with AMP2026 – presentA runway walk for the Booster K1 learned with adversarial motion priors.
RFCP, Robot Fashion Catwalk Policy2026-08 – 2026-09Our Ghost Trial 04 entry. A G1 walks the runway, turns, poses and walks back, by fine-tuning SONIC.
K1 Heel Shoes2026-09 – present3D-printed heel shoes for the Booster K1.
K1 Heel Walking2026-10 – presentGetting the K1 walking policies to walk in the 50 mm heels. Sim for now.
K1 Loco-Teleop2026-10 – presentA walking policy for the Booster K1 that drives only the legs, so the arms and head stay free for teleoperation.
K1 Cartwheel2026-10 – presentTeaching the Booster K1 a cartwheel from a video. It works in simulation; on the robot it still falls.
Mi-RoL2026A multi-tier LLM framework for empathetic conversation on the Pepper robot, grounded in Motivational Interviewing. Tested in a 40-participant study.
Structured prompting and knowledge bias in VLMs2026 – 2026-05Can structured prompting overcome knowledge bias in vision-language models? Four open-source VLMs, five prompting strategies, the VLMBias benchmark.
Microduck Arena Lab2026-10Faster running gaits for Pollen Robotics' Microduck on three simulated race courses, fine-tuned with PPO and chosen by the Arena's five-seed score. #1 on the roller sprint leaderboard.

Technical skills

Robots
Booster K1 (lab fleet of six), Unitree G1, Pepper
Robotics / RL
Isaac Lab/Sim, mjlab, MuJoCo, ROS 2, PPO/RSL-RL, AMP, BeyondMimic, NVIDIA SONIC, GVHMR, GMR, sim-to-real
ML / AI
PyTorch, Hugging Face, vLLM, ONNX Runtime, OpenCV, scikit-learn, TensorFlow, LLMs, VLMs, RAG, FAISS
Languages
Python, SQL, JavaScript/TypeScript, Bash, R, LaTeX
Systems
Linux, Git, Docker, conda, CMake, CUDA, H100 GPU servers, FastAPI, PostgreSQL, React, Node.js, CAD, 3D printing

Honours

Dean's ListSpring 2026Kent State University
Merit scholarship (×5)Vilnius University
Exchange grantsISEP (×2), Erasmus+

Teaching

Workshop on humanoid imitation learning and sim-to-real transfer (4 hours)Kent State University, university-wide
Three workshops for lab membersAdvanced Telerobotics Research LabTwo on robot learning, one on reinforcement learning.