KocDigital (2021 – 2023) Data ScientistÂ
• Assisted in an R&D (Reinforcement Learning) project on task allocation of a multi-robot system.
• Implemented Conservative Q-learning (CQN) and Soft Actor Critic (SAC) algorithms for credit pricing problem in banking sector.
• Developed a wildfire prediction model for Turkey’s Ministry of Agriculture and Forestry. The project was supported by World Economic Forum and presented in Davos.
• Worked on the DaaS (Data as a Service) project, in which I was responsible for building prediction models for inflation, interest rate, and currency rates.
DenizBank Head Office (2019 – 2021) Data ScientistÂ
• Developed collection (self-cure and roll forward) models for overdraft and mortgage credits.
• Collaborated with BDDK (Banking Regulation and Supervision Agency) on the EWS (early warning signal) project.
• Tested & developed probability of default (PD) models of the bank.
• Worked on limit optimization for general purpose loan (GPL).
Labs & Projects:
Omdena | Graph-Based Learning (2019)
• Used graph-based learning to find the best route for families to get reunited aftermath of expected Istanbul earthquake.
Koc University & Inzva | Computer Vision for Autonomous Driving (2020)
• AI labs joint program where the following topics were covered under the supervision of Asst. Prof. Fatma Guney: Object detection, visual object tracking, semantic segmentation.
Inzva | Using Decision Transformer in Drone Environment (2021)
• We have trained the Decision Transformer agent in Pybullet’s drone simulator environment and compared the results with the model-free RL benchmarks.