Kaan Pekel Email and Phone Number

• Designed and implemented a scoring algorithm using Python and SQL for 200K+ sellers at Trendyol, Turkey’s first decacorn and largest e-commerce platform, helping customers identify trustworthy sellers and motivating seller improvement through Looker dashboards, resulting in increased customer NPS and seller NPS• Managed production workflows using Airflow and Cron, overseeing 50+ ETL processes and leveraging BigQuery and Apache Kafka for real-time data communication, ensuring seamless model execution and maintenance in production.• Led strategy for sales forecasting models, fraud detection algorithms, seller clustering and coupon recommendation algorithms, defining the approach and overseeing project execution while delegating operational tasks.• Optimized the Buy Box algorithm, improving conversion rates and click-through-rates through conducting A/B tests on segmented user groups across web and mobile for 100M+ unique product-seller combinations in the marketplace.

• Developed a brand and category propensity algorithm to identify users' interests and recommend favorite components, delivering personalized content that significantly increased conversion rates and reduced bounce rates.• Designed and implemented a clustering and homepage sorting algorithm for Trendyol’s main app and meal/instant delivery app, delivering personalized components, enabling effective targeting and enhancing personalization for 40M+ customers.• Led the migration to GitLab, overseeing repository relocation and managing code reviews to ensure best practices in version control.• Promoted agile methodologies and fostered collaborative problem-solving within the team, enhancing development cycles and improving stakeholder management. Created documentation for each project, streamlining onboarding for new team members.

Modeled a complex disease-symptom network, developing a classifier framework and simulating failure spread using Java to predict outbreak risks. Analyzed and visualized experiment results using R (tidyverse), creating dashboards in Tableau and comparing its performance against state-of-the-art classification algorithms.Article published in Computers & Operations Research Journal:https://www.sciencedirect.com/science/article/abs/pii/S0305054824000704

• Created dashboards & reports using Google Analytics & Ads, Power BI and Excel to track and improve customer retention for Insider, an AI-native Omnichannel Experience and Customer Engagement Platform with 1200+ partners, one of the first unicorns in Turkey. • Conducted performance analysis of predictive marketing algorithms, including predictive ad audiences and recommendation engines, and developed digital marketing strategies.

I was the teaching asisstant for the following courses:- INDR 252, Applied Statistics, Fall 2018.- MATH 102, Calculus, Spring 2018.- MATH 101, Finite Mathematics, Fall 2017.