Ugur Kaplan

In collaboration with ”Istanbul Technical University - Eurasia Institute of Earth Sciences Ocean and Atmospheric Sciences Group” in the context of Time Series Forecasting: ∗ I Trained various LSTM-based models for Seq2Seq prediction. ∗ Applied iterative Gaussian Processes to time-series data for one-step ahead prediction with uncertainty. ∗ Currently using state-space models with (Extended) Kalman and Particle Filters for time-series forecasting with uncertainty.

* Conducted statistical analysis on patient data and developed multiple machine learning models. Used MLFlow and later on, Weights & Biases for experiment tracking. * Deployed machine learning models to on-premise servers of the infertility clinic using Ray Serve. * I designed classes according to the OOP design patterns (GoF), documented the code, written unit and integration tests. * I created Docker images for test and development environments.