I'm a data scientist.
ChatGPT-4o’s Perspective of What I Do:
“Ender connects the dots between data and human behavior — turning Python into clarity, dashboards into decisions, and pull-ups into balance.”
• 5+ years of experience applying analytics and machine learning to solve business challenges.
• Led projects in product segmentation, sales forecasting, and recommender systems.
• Ranked 1st in class at Sabancı University (BSc GPA 3.93 / MSc GPA 3.95), where my Master’s thesis focused on data-driven inventory and production decision-making for a major tire manufacturer.
• Skilled in Python, R, SQL, advanced statistical methods, and data visualization.
• Passionate about calisthenics, jump roping, and maintaining a healthy lifestyle.
• Deeply interested in reading about cognitive biases, logical fallacies, and decision-making.
• Led analytics projects by analyzing large-scale search logs and user behavior using Python, SQL, clustering, and statistical modeling to uncover trends, track key metrics (e.g., DAU, CTR), and guide data-driven improvements to search quality through experimentation and cross-functional collaboration.
​
• Designed and led the SBS-NPS project to evaluate search quality through statistical analysis of human-labeled comparisons, transforming qualitative judgments into a reliable metric for experimentation, evaluation, and product improvement.
​
• Built automated data pipelines and dashboards to collect, validate, and visualize key evaluation metrics, ensuring data quality, transparency, and actionable feedback across the search quality workflow.
​
• Developed machine learning models to personalize agent-customer interactions using Python, SQL, and R, leveraging large-scale behavioral data to predict preferences and rank call recommendations.
• Applied data mining, feature engineering, A/B testing, and statistical bootstrapping to enhance model accuracy and robustness in a real-time production environment.
• Collaborated with research scientists and engineers to refine recommendation logic, contributing to a 5% increase in sales revenue.
• Applied the CRISP-DM framework to lead end-to-end analysis of inventory and sales data, aligning stock keeping unit (SKU) strategy with production and business goals using Python.
• Designed and implemented a scoring-based SKU prioritization system using unsupervised learning techniques to support production planning.
• Performed scenario and counterfactual analysis, along with simulation studies, to evaluate the impact of prioritization strategies, enabling data-driven production decisions and reducing excess inventory by 20%.
• Designed a hybrid forecasting framework that combined machine learning models with human expertise to improve inventory planning and forecast accuracy across product lines.
​
• Conducted empirical research on human intervention in algorithmic forecasting, identifying when to rely on automation, human judgment, or a human-in-the-loop approach.
​
• Optimized forecasting performance by segmenting products based on predictability and business context, leading to more accurate inventory decisions and reduced planning errors.
• Designed and deployed a data-driven web interface with real-time dashboards to support strategic planning and sales decision-making. View Project Report
​
• Developed sales forecasting models using historical demand data and applied predictive techniques, including power, logarithmic, exponential, and polynomial regression, improving forecast accuracy by 10%.
​
• Applied association rule mining (market basket analysis) to identify frequently co-purchased products, informing van loading strategies that reduced logistics costs and improved retailer satisfaction.​