Mehmet Ozan Baykan

Analyzed model-generated responses and curated training datasets focused on advanced topics in mathematics, statistics, data science, and machine learning to support Reinforcement Learning from Human Feedback (RLHF) training pipelines for state-of-the-art reasoning LLMs.

Development of Unscented and Extended Kalman Filters to quantify uncertainty for attitude estimation in Python.

- Investigated the biophysical characterization of protein dynamics underlying TMP resistance among E. Coli DHFR mutants by physics-informed statistical analysis of Molecular Dynamics (MD) simulation data using Python (ProDy, SciPy, scikit-learn) and VMD. - Implemented regression (GLM, Bayesian regression), dimensionality-reduction (PCA, UMAP) and time-series (RMSD) methods to extract statistical insight and inform causal analysis.

- Collaborated with a multidisciplinary team of bachelor students in an independent research project in therapeutic applications of bacterial neurotoxins for the iGEM Competition. - Responsible for mathematical and computational modeling and conducting wet-lab experiments.

- Tutor for CS115 - Introduction to Programming in Pythonç - Organized and helped students in weekly online coding sessions in coordination with professors and teaching assistants.