Istanbul, Türkiye
I’ve graduated from Yıldız Technical University with a Bachelor of Statistics. Data adventure has started in 2021 for me. Since then, I’ve been building projects and improving myself at the field of data science. Not only with data science, but I’ve been also trying to improve myself with technology around which helps a lot to understand what challenges are. I also deal with time series’ forecasting by using machine learning techniques. While these actions are going on, I used a few programming languages to build my own projects. Kind of programming languages like Python and R have helped me on this challenging road. Programming Skills Include; •Database programming such as MsSql •Python programming with knowledge of lots of libraries such as Pandas, NumPy, Scikit-learn, Keras, TensorFlow, Matplotlib, Seaborn, etc. •R programming with the high level of knowledge for performing at data science field by using packages such as dplyr, sqldf, ggplot2, etc. •Business Intelligence tools such as PowerBI with knowledge of connection between Database servers and PowerBI.
• Propensity Modelling | XGBoost, LightGBM, Bayesian Optimization Developed tree-based machine learning models to predict customer propensity to accept insurance offers using historical quote data. Employed Bayesian optimization for hyperparameter tuning and implemented a custom iterative feature selection strategy based on cumulative information gain, retaining features covering 99.5% of total gain across successive model refinements.
• Earthquake Claim Risk Modeling | Bayesian & Monte Carlo Methods Designed a probabilistic risk model to estimate expected claims from a future Marmara earthquake. Applied Bayesian inference with structural exposure variables and historical data from the Kahramanmaraş earthquake. Built a Monte Carlo simulation using geospatial and policyholder data to quantify claim distributions. • MTPL Pricing Model | GLM, Radar, Emblem Prepared and integrated MTPL data for risk modeling using Radar, Emblem, and Python; developed GLM and GBM models and supported end-to-end system integration for pricing automation. • Claim Inflation Forecasting Automation | Python, ARIMAX, LSTM, RNN Redundant Excel-based process was automated using advanced time series models in Python. Improved forecasting accuracy and scalability by implementing ARIMAX and deep learning models (LSTM, RNN) with macroeconomic indicators. Resulted in a more robust and maintainable workflow.
• Revenue Forecast Automation | Survival Analysis & Weibull Modeling Led end-to-end development of an automated revenue forecasting tool using survival analysis and Weibull distribution. Designed and implemented the model in Python with data integration via Zoho Analytics, leveraging CRM data for probabilistic forecasting. The solution significantly improved monthly and annual revenue predictions after a successful pilot rollout.
• Stock Market Analysis & Price Prediction | Association Rules, Clustering, LSTM Utilized association rule mining and clustering to explore stock exchange patterns. Developed LSTM models in TensorFlow and Keras, incorporating exogenous variables such as foreign exchange forecasts. Engineered features from historical prices, trading volumes, and macroeconomic indicators to improve model performance for NASDAQ and XU100 indices. • Time Series Forecasting & Volatility Modeling Implemented SES, Holt’s Method, ARIMA (with grid search parameter tuning), ARCH, and GARCH models to analyze market trends and volatility. Evaluated ARIMA performance on both stationary and non-stationary data, leveraging ARCH/GARCH for precise volatility estimation in stock returns. • Monte Carlo Simulation for Financial Risk Modeling Applied Monte Carlo simulations with stochastic differential equations to simulate stock price dynamics and quantify risk across multiple market conditions.