I am a machine learning engineer and a Ph.D. Candidate studying Machine Learning in the Climate & Hydrometeorology domains. I have participated in multidisciplinary teams for climate, law, and medicine projects.

I have experience utilizing high-performance computing (HPC) clusters, distributed computing, and parallelization. Although I prefer Python, Bash, and
SQL, I also have experience in R, C#, and no-SQL datasets.

I am a registered member of the Ankara branch of the Turkish Chamber of Civil Engineers and The Turkish Foundation for Combating Soil Erosion for Reforestation and the Protection of Natural Habitats (TEMA).

I am the human of three rescued cats (Bal, Misket, and Bulut) and one budgie
(Maşuk). Also, I would like to mention Duman, a semi-stray cat (who sometimes
visited houses that let him in and refused to be permanently homed) whom I
cared for when he was sick but sadly passed away.

Further Details

I believe in explainable, trustworthy, reliable, and robust AI solutions. To this end, I enjoy participating in novel projects and publishing novel works.

In many real-world applications, the dataset is biased against some classes (e.g., some cases are underrepresented in medical imaging datasets). I enjoy wielding imbalance mitigation tools (e.g., class & sample weights, tailored loss functions, over & undersampling) and taking on these datasets.

Similarly, similar to most humans, I would like to know how and why my models generate their predictions. Knowing the reasoning at the model level (i.e., parameter importance and casual relations) would increase computational efficiency in time and cost. Likewise, knowing the reasoning at the instance level would enable me to correct my model when it mispredicts.

Lastly, I think a good machine learning model should provide information about how sure it is about particular predictions. This way, the reliability and trustworthiness of the models will be increased. I utilize ensemble and Bayesian approaches (e.g., Monte Carlo Dropout) for uncertainty quantization.