Istanbul, Türkiye
AI developer with expertise in Fintech
AI Department
Infrastructure Applications Department
Over the past 3 years, I have create a cutting-edge stock exchange trading terminal operates on the cloud with real-time market insights and seamless execution capabilities in milliseconds. I gained a competitive edge in the fast-paced financial markets by developing AI-powered trading robots that leverage deep learning algorithms and makes data-driven decisions.
Developed AI backend of MariksIQ product •.Net Core Class Library: AI.Data I developed a core ETL library in C# for time series data processing in the context of machine learning. This library encompasses normalization techniques, windowing, splitting, and reshaping functionalities, and it is compatible with financial time series data. Additionally, the library includes unit tests for quality assurance. •WPF Class Library: Alpx.UI.DX When a desktop application utilizing WPF technology undergoes an update involving a substantial volume of data, the process of redrawing each data update becomes time-consuming, leading to a slowdown in the program's visual performance (less than 30 FPS). I conducted a benchmark that compared DirectX technology to the existing technology when updating price data in a 100-cell table. Subsequently, I developed a library aimed at enhancing the UI update performance of DrawingVisual in WPF by harnessing DirectX technologies. This innovation results in a substantial improvement in the drawing speed of desktop program UI, achieving performance boosts ranging from x6 to x12 times. This achievement incorporates four distinct optimization techniques and efficiently utilizes computational power. •Python Microservice: ANN Price Predictor I analyzed, debugged, fixed and improved a private TensorFlow deep learning price prediction model, resulting in a remarkable increase in model prediction accuracy from 53% to 61%. My efforts extended to researching and enhancing the LSTM layer to align it with the appropriate data shape and optimize performance. Furthermore, I generated visualizing reports using Matplotlib to provide insightful visual representations of the model's performance and predictions.
During my time as a college student, I participated in the United States' Work and Travel program and I had the chance to live in Delaware for a rewarding period of 4 months. During this time, I worked at a hotel with a captivating ocean view, affording me a unique opportunity to closely observe and understand the intricacies of tourism management and operation. This experience significantly contributed to my personal growth and enhanced my language skills, I achieved notable english fluency through daily conversations