Michal Wilinski

I am a final-year B.Sc. student in Artificial Intelligence at Poznan University of Technology and a Research Fellow at the Auton Lab, Robotics Institute, Carnegie Mellon University. My research focuses on foundation models for time series data, with applications in healthcare.

Currently, I’m investigating benchmarking and post-training methods for LLM agents in time series, while also exploring the effectiveness of time series foundation models in real-world medical applications. I previously participated in the Robotics Institute Summer Scholar (RISS) program at CMU, where I lead the project on mechanistic interpretability of Time Series Foundation Models. I also contributed to projects on long-context Time Series Foundation Models and implicit reasoning in deep time series forecasting. I was lucky to work for a bit with Team Chiron in DARPA Triage Challenge.

At the Institute of Robotics & Machine Intelligence (Poznan University of Technology), I work on haptic perception and locomotion learning in collaboration with Prof. Jan Peters’ IAS group at TU Darmstadt. I’ve also conducted research on SLAM algorithms and localization for (1) space environments in partnership with KP Labs and the European Space Agency and (2) industrial environments in collaboration with Samsung.

Research Interests

• Foundation Models
• Representation Learning in Large Models
• Reinforcement Learning
• Human-AI Interaction

Education

• B.Sc. in Artificial Intelligence, Poznan University of Technology (2021-2025)
  • GPA: 4.58/5.0
  • University President’s Merit Scholar
  • Expected Final Grade: 5.0/5.0

Recent Publications

My research has been published at venues including NeurIPS Workshops and CoRL. For a complete list, please visit my publications page.

Contact me if you’re interested in collaboration or have any questions about my research.